{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b2f8545bfe2a",
   "metadata": {},
   "source": [
    "# Logistic regression analyses: linear and cubic models\n",
    "\n",
    "The same lung SUV data are analysed on both the **plain** and **log-transformed** predictor scales using three regression models:\n",
    "\n",
    "1. [Standard (linear) logistic regression](#2-standard-linear-logistic-regression)\n",
    "2. [Cubic unconstrained (non-monotonic) logistic regression](#3-cubic-unconstrained-non-monotonic-logistic-regression)\n",
    "3. [Cubic constrained (monotonic) logistic regression](#4-cubic-constrained-monotonic-logistic-regression)\n",
    "\n",
    "\n",
    "Each section discussed a separate logistic model.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8f6904454fdc",
   "metadata": {},
   "source": [
    "## 1. Shared setup and data\n",
    "\n",
    "The data-loading and predictor-transformation steps are common to all three models. Model-specific figures are written to separate subdirectories below `../results/logit_models/`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "737bb31e0d5f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from pathlib import Path\n",
    "import sys\n",
    "\n",
    "project_root = Path.cwd().parent   # when notebook is in project/notebooks/\n",
    "src_dir = project_root / \"src\"\n",
    "\n",
    "if str(src_dir) not in sys.path: sys.path.insert(0, str(src_dir))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "3b867066e59d",
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "import pandas as pd\n",
    "\n",
    "# our libs\n",
    "from irae_risk import data\n",
    "from irae_risk import logistic\n",
    "\n",
    "np.set_printoptions(precision=16)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "86f652a20401",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_path = Path(\"../data/raw\")\n",
    "results_root = Path(\"../results/logit_all_models\")\n",
    "results_root.mkdir(parents=True, exist_ok=True)\n",
    "\n",
    "df = data.load_mat_data_to_dataframe(\n",
    "    suv_filename=data_path / \"suv_percentilesSLOthenUWM.mat\",\n",
    "    flags_filename=data_path / \"flags_combined.mat\",\n",
    "    nr_patients=58,\n",
    ")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "5956c35b9f37",
   "metadata": {},
   "outputs": [],
   "source": [
    "# get data\n",
    "perc = 95\n",
    "organ = \"lung\"\n",
    "\n",
    "x0, y = data.get_data_from_dataframe(df, organ, perc)\n",
    "\n",
    "scales = [\"plain\", \"log\"]\n",
    "xs = [x0, np.log(x0)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f8e352bec94c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Two-sided 95% confidence intervals used throughout the notebook\n",
    "alpha = 0.05\n",
    "probs = [alpha / 2, 1 - alpha / 2]\n",
    "\n",
    "# Reproducibility: pass this seed to every stochastic calculation.\n",
    "random_seed = logistic.DEFAULT_RANDOM_SEED"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "56c8948cbea0",
   "metadata": {},
   "source": [
    "## 2. Standard (linear) logistic regression\n",
    "\n",
    "The log-odds are linear in the selected predictor scale,\n",
    "\n",
    "$$\n",
    "\\eta(t) = \\beta_0 + \\beta_1 t,\n",
    "\\qquad\n",
    "p(t) = \\operatorname{expit}\\!\\left(\\eta(t)\\right).\n",
    "$$\n",
    "\n",
    "The model is unconstrained (`degree=1`, `mono=False`). The analysis includes the fit, asymptotic parameter covariance and confidence intervals, goodness-of-fit measures, and normal/delta confidence bands for the fitted risk curve.\n",
    "\n",
    "References: [scikit-learn logistic regression documentation](https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression) and [discussion of the optimized objective](https://stats.stackexchange.com/questions/186830/what-is-scikit-learns-logisticregression-minimizing).\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "39cbb1957ccb",
   "metadata": {},
   "outputs": [],
   "source": [
    "results_path = results_root / \"linear\"\n",
    "results_path.mkdir(parents=True, exist_ok=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "95e977ed6c50",
   "metadata": {},
   "source": [
    "### 2.1 Fit and parameter uncertainty\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "774797739c35",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "thetas = array([[-11.68387204638564 ,   5.686965443671553],\n",
      "       [ -7.527409035255565,  10.73815986880034 ]])\n"
     ]
    }
   ],
   "source": [
    "# unconstrained linear logistic-regression model\n",
    "logit = logistic.LogisticPolyRegression(degree=1, mono=False)\n",
    "\n",
    "# fits\n",
    "fit_results = [logit.fit(x, y, method=\"local\", seed=random_seed) for x in xs]\n",
    "if not all(result[\"success\"] for result in fit_results):\n",
    "    raise RuntimeError(\"At least one linear logistic-regression fit failed.\")\n",
    "\n",
    "thetas = np.array([result[\"theta\"] for result in fit_results])\n",
    "print(f\"{thetas = }\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "82a3535b87ed",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cov_thetas = array([[[12.323037015518926 , -6.7129191282761775],\n",
      "        [-6.7129191282761775,  3.8187839258530176]],\n",
      "\n",
      "       [[ 5.0502362164436665, -7.934551247510305 ],\n",
      "        [-7.934551247510305 , 14.00597781508227  ]]])\n"
     ]
    }
   ],
   "source": [
    "# asymptotic covariance matrix of parameters\n",
    "cov_thetas = np.array([logit.get_cov(x, y, theta) for x, theta in zip(xs, thetas)])\n",
    "print(f\"{cov_thetas = }\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a4a84bad7b8c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "theta_CIs = array([[[-18.56416558961513  ,   1.8568608502834008],\n",
      "        [ -4.803578503156151 ,   9.517070037059703 ]],\n",
      "\n",
      "       [[-11.931983300748936 ,   3.4030806585808593],\n",
      "        [ -3.1228347697621963,  18.073239079019817 ]]])\n"
     ]
    }
   ],
   "source": [
    "# CI of params (assuming asymptotic distr of parameters)\n",
    "theta_CIs = np.array([\n",
    "    logit.get_theta_quantiles_normal(probs, theta, cov_theta)\n",
    "    for theta, cov_theta in zip(thetas, cov_thetas)\n",
    "])\n",
    "\n",
    "print(f\"{theta_CIs = }\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "b2e5269ad11c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "('scale', 'coef')",
         "rawType": "object",
         "type": "unknown"
        },
        {
         "name": "value",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "LCL",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "UCL",
         "rawType": "float64",
         "type": "float"
        }
       ],
       "ref": "34b0cd91-26a1-48de-a260-1685c058bb1b",
       "rows": [
        [
         "('plain', 'b0')",
         "-11.68387204638564",
         "-18.56416558961513",
         "-4.803578503156151"
        ],
        [
         "('plain', 'b1')",
         "5.686965443671553",
         "1.8568608502834008",
         "9.517070037059703"
        ],
        [
         "('log', 'b0')",
         "-7.527409035255565",
         "-11.931983300748936",
         "-3.1228347697621963"
        ],
        [
         "('log', 'b1')",
         "10.73815986880034",
         "3.4030806585808593",
         "18.073239079019817"
        ]
       ],
       "shape": {
        "columns": 3,
        "rows": 4
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>value</th>\n",
       "      <th>LCL</th>\n",
       "      <th>UCL</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scale</th>\n",
       "      <th>coef</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">plain</th>\n",
       "      <th>b0</th>\n",
       "      <td>-11.683872</td>\n",
       "      <td>-18.564166</td>\n",
       "      <td>-4.803579</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b1</th>\n",
       "      <td>5.686965</td>\n",
       "      <td>1.856861</td>\n",
       "      <td>9.517070</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"2\" valign=\"top\">log</th>\n",
       "      <th>b0</th>\n",
       "      <td>-7.527409</td>\n",
       "      <td>-11.931983</td>\n",
       "      <td>-3.122835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b1</th>\n",
       "      <td>10.738160</td>\n",
       "      <td>3.403081</td>\n",
       "      <td>18.073239</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                value        LCL        UCL\n",
       "scale coef                                 \n",
       "plain b0   -11.683872 -18.564166  -4.803579\n",
       "      b1     5.686965   1.856861   9.517070\n",
       "log   b0    -7.527409 -11.931983  -3.122835\n",
       "      b1    10.738160   3.403081  18.073239"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# present table of results\n",
    "d = thetas.shape[-1]\n",
    "index = pd.MultiIndex.from_tuples([(lab, f\"b{i}\") for lab in scales for i in range(d)], names = ['scale', 'coef'])\n",
    "pd.DataFrame({\"value\": thetas.flatten(), \"LCL\": theta_CIs[:,0].flatten(), \"UCL\": theta_CIs[:,1].flatten()}, index = index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "5fc72adc9c57",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "scale",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "LLF",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "AIC",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "BIC",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "A",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "deviance",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "p-value(deviance_bootstrap)",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "deviance_bootstrap_samples",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "n",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "k",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "dof",
         "rawType": "int64",
         "type": "integer"
        }
       ],
       "ref": "939599f3-8248-418e-b295-7d8f8c5ecfd3",
       "rows": [
        [
         "plain",
         "-7.394195051363893",
         "18.78839010272779",
         "22.909276123820625",
         "0.9655172413793104",
         "14.788390102727787",
         "0.3956043956043956",
         "1000",
         "58",
         "2",
         "56"
        ],
        [
         "log",
         "-6.795375128439737",
         "17.590750256879474",
         "21.71163627797231",
         "0.9655172413793104",
         "13.590750256879474",
         "0.39760239760239763",
         "1000",
         "58",
         "2",
         "56"
        ]
       ],
       "shape": {
        "columns": 10,
        "rows": 2
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>LLF</th>\n",
       "      <th>AIC</th>\n",
       "      <th>BIC</th>\n",
       "      <th>A</th>\n",
       "      <th>deviance</th>\n",
       "      <th>p-value(deviance_bootstrap)</th>\n",
       "      <th>deviance_bootstrap_samples</th>\n",
       "      <th>n</th>\n",
       "      <th>k</th>\n",
       "      <th>dof</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scale</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>-7.394195</td>\n",
       "      <td>18.78839</td>\n",
       "      <td>22.909276</td>\n",
       "      <td>0.965517</td>\n",
       "      <td>14.78839</td>\n",
       "      <td>0.395604</td>\n",
       "      <td>1000</td>\n",
       "      <td>58</td>\n",
       "      <td>2</td>\n",
       "      <td>56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>log</th>\n",
       "      <td>-6.795375</td>\n",
       "      <td>17.59075</td>\n",
       "      <td>21.711636</td>\n",
       "      <td>0.965517</td>\n",
       "      <td>13.59075</td>\n",
       "      <td>0.397602</td>\n",
       "      <td>1000</td>\n",
       "      <td>58</td>\n",
       "      <td>2</td>\n",
       "      <td>56</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            LLF       AIC        BIC         A  deviance  \\\n",
       "scale                                                      \n",
       "plain -7.394195  18.78839  22.909276  0.965517  14.78839   \n",
       "log   -6.795375  17.59075  21.711636  0.965517  13.59075   \n",
       "\n",
       "       p-value(deviance_bootstrap)  deviance_bootstrap_samples   n  k  dof  \n",
       "scale                                                                       \n",
       "plain                     0.395604                        1000  58  2   56  \n",
       "log                       0.397602                        1000  58  2   56  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# goodness of fit measures\n",
    "index = pd.Index(scales, name = 'scale')\n",
    "pd.DataFrame(\n",
    "    [logit.goodness_of_fit(x, y, theta, bootstrap_seed=random_seed)\n",
    "     for x, theta in zip(xs, thetas)],\n",
    "    index=index,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32deb7403fd0",
   "metadata": {},
   "source": [
    "### 2.2 Fitted risk curve and confidence bands\n",
    "\n",
    "**Coverage note.** The term *confidence band* is used in its standard graphical sense. The displayed limits are pointwise confidence intervals evaluated separately at each predictor value; they do not provide simultaneous coverage of the complete fitted curve.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "5630ea114f77",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plotting results\n",
    "fig, axs = plt.subplots(ncols = len(scales), figsize = (6*len(scales), 5))\n",
    "\n",
    "quantile_methods = {\n",
    "    \"normal\": lambda x, p, theta, cov: logit.get_model_quantiles_normal(\n",
    "        x, p, theta, cov, seed=random_seed\n",
    "    ),\n",
    "    \"delta\": logit.get_model_quantiles_delta,\n",
    "}\n",
    "\n",
    "for ax, scale, x, theta, cov_theta in zip(axs, scales, xs, thetas, cov_thetas):\n",
    "\n",
    "    ax.set_title(f\"cond. prob. for AE: scale {scale}\")\n",
    "\n",
    "    if scale == \"plain\":\n",
    "        xlab = r\"$x = max_{visit} SUV(visit, p)$\"\n",
    "    else:\n",
    "        xlab = r\"$x = log(max_{visit} SUV(visit, p))$\"\n",
    "    \n",
    "    ax.set_xlabel(xlab)\n",
    "    ax.set_ylabel(\"P(AE|X = x)\")\n",
    "\n",
    "    # plotting data points\n",
    "    for i, group_label in enumerate([\"NC\", \"AE\"]):\n",
    "        ax.scatter(x[y == i], y[y == i], label=group_label)\n",
    "\n",
    "    # plotting fitted model \n",
    "    xp = np.linspace(min(x), max(x), 100)\n",
    "    yp = logit.model(xp, theta)\n",
    "\n",
    "    ax.plot(xp, yp, label = \"logistic reg\")\n",
    "\n",
    "    for ci_label, color in zip([\"normal\", \"delta\"], [\"red\", \"green\"]):\n",
    "        # get result of model quantiles at probs\n",
    "        # theta and cov_theta are obtained via MLE method\n",
    "        res = quantile_methods[ci_label](xp, probs, theta, cov_theta)\n",
    "\n",
    "        # make plot of quantiles\n",
    "        ax.fill_between(xp, *res, color=color, alpha=0.5, label=f\"CI:{ci_label}\")\n",
    "\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "fig.legend(handles, labels, bbox_to_anchor=(1.02, 0.5), loc='center right')\n",
    "\n",
    "fig.savefig(results_path / \"logit_linear_fit.pdf\", bbox_inches=\"tight\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "f9fb1cb06129",
   "metadata": {},
   "outputs": [],
   "source": [
    "linear_results = {\n",
    "    \"model\": logit,\n",
    "    \"fit_results\": fit_results,\n",
    "    \"parameters\": thetas,\n",
    "    \"covariances\": cov_thetas,\n",
    "    \"parameter_ci_wald\": theta_CIs,\n",
    "}\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4971dd5fff15",
   "metadata": {},
   "source": [
    "## 3. Cubic unconstrained (non-monotonic) logistic regression\n",
    "\n",
    "The log-odds are represented by an unconstrained cubic polynomial,\n",
    "\n",
    "$$\n",
    "\\eta(t) = \\beta_0 + \\beta_1 t + \\beta_2 t^2 + \\beta_3 t^3.\n",
    "$$\n",
    "\n",
    "Because no derivative constraint is imposed (`degree=3`, `mono=False`), the fitted probability curve may be non-monotonic. The same uncertainty and goodness-of-fit calculations used for the linear model are repeated here.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "98a60d53d9fc",
   "metadata": {},
   "outputs": [],
   "source": [
    "results_path = results_root / \"cubic_unconstrained\"\n",
    "results_path.mkdir(parents=True, exist_ok=True)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f50ae75ed9fc",
   "metadata": {},
   "source": [
    "### 3.1 Fit and theta-parameter uncertainty\n",
    "\n",
    "The calculations below are written in terms of the fitted parameters $\\theta$. For the unconstrained cubic model the parameterization is the identity, $\\beta=\\theta$, so these intervals are also intervals for the polynomial coefficients $\\beta$.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "430ae4924260",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "thetas = array([[-197.43456924470524,  248.4905000484067 ,  -98.59976324737492,\n",
      "          12.18035307606018],\n",
      "       [ -75.33868079212766,  287.5173164621367 , -339.6550928412925 ,\n",
      "         124.51434906316405]])\n"
     ]
    }
   ],
   "source": [
    "# unconstrained cubic logistic-regression model\n",
    "logit = logistic.LogisticPolyRegression(degree=3, mono=False)\n",
    "\n",
    "# fits\n",
    "fit_results = [logit.fit(x, y, method=\"local\", seed=random_seed) for x in xs]\n",
    "if not all(result[\"success\"] for result in fit_results):\n",
    "    raise RuntimeError(\"At least one cubic logistic-regression fit failed.\")\n",
    "\n",
    "thetas = np.array([result[\"theta\"] for result in fit_results])\n",
    "print(f\"{thetas = }\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "10268adb7723",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cov_thetas = array([[[ 10939.684523567224   , -14010.646252568109   ,\n",
      "           5653.962064031723   ,   -708.3678994726915  ],\n",
      "        [-14010.646252568109   ,  18013.760997744266   ,\n",
      "          -7300.014221918021   ,    918.9939676648023  ],\n",
      "        [  5653.962064031723   ,  -7300.014221918021   ,\n",
      "           2973.1108397867397  ,   -376.64839161858305 ],\n",
      "        [  -708.3678994726915  ,    918.9939676648023  ,\n",
      "           -376.64839161858305 ,     48.140910244592945]],\n",
      "\n",
      "       [[  1998.703486237727   ,  -7705.433239899273   ,\n",
      "           9196.92321106423    ,  -3384.606945444059   ],\n",
      "        [ -7705.433239899273   ,  29940.973099065315   ,\n",
      "         -35991.86768534103    ,  13336.386966233815   ],\n",
      "        [  9196.92321106423    , -35991.86768534103    ,\n",
      "          43591.818688455074   , -16286.649984027337   ],\n",
      "        [ -3384.606945444059   ,  13336.386966233815   ,\n",
      "         -16286.649984027337   ,   6149.469270433855   ]]])\n"
     ]
    }
   ],
   "source": [
    "# asymptotic covariance matrix of parameters\n",
    "cov_thetas = np.array([logit.get_cov(x, y, theta) for x, theta in zip(xs, thetas)])\n",
    "print(f\"{cov_thetas = }\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "233e3b495e13",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "theta_CIs = array([[[-402.4329780401599  ,  -14.566758126806093,\n",
      "         -205.4692297024859  ,   -1.418592629226529],\n",
      "        [   7.56383955074935 ,  511.54775822361944 ,\n",
      "            8.269703207736043,   25.77929878134689 ]],\n",
      "\n",
      "       [[-162.96251965553296 ,  -51.62426932699947 ,\n",
      "         -748.8691888972139  ,  -29.18318823491893 ],\n",
      "        [  12.285158071277635,  626.6589022512728  ,\n",
      "           69.55900321462877 ,  278.211886361247   ]]])\n"
     ]
    }
   ],
   "source": [
    "# CI of params (assuming asymptotic distr of parameters)\n",
    "theta_CIs = np.array([\n",
    "    logit.get_theta_quantiles_normal(probs, theta, cov_theta)\n",
    "    for theta, cov_theta in zip(thetas, cov_thetas)\n",
    "])\n",
    "print(f\"{theta_CIs = }\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "18a1b339bcd7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "('scale', 'coef')",
         "rawType": "object",
         "type": "unknown"
        },
        {
         "name": "value",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "LCL",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "UCL",
         "rawType": "float64",
         "type": "float"
        }
       ],
       "ref": "8a5fddf8-b277-469e-ab6a-08d9016626ec",
       "rows": [
        [
         "('plain', 'b0')",
         "-197.43456924470524",
         "-402.4329780401599",
         "7.56383955074935"
        ],
        [
         "('plain', 'b1')",
         "248.4905000484067",
         "-14.566758126806093",
         "511.54775822361944"
        ],
        [
         "('plain', 'b2')",
         "-98.59976324737492",
         "-205.4692297024859",
         "8.269703207736043"
        ],
        [
         "('plain', 'b3')",
         "12.18035307606018",
         "-1.418592629226529",
         "25.77929878134689"
        ],
        [
         "('log', 'b0')",
         "-75.33868079212766",
         "-162.96251965553296",
         "12.285158071277635"
        ],
        [
         "('log', 'b1')",
         "287.5173164621367",
         "-51.62426932699947",
         "626.6589022512728"
        ],
        [
         "('log', 'b2')",
         "-339.6550928412925",
         "-748.8691888972139",
         "69.55900321462877"
        ],
        [
         "('log', 'b3')",
         "124.51434906316405",
         "-29.18318823491893",
         "278.211886361247"
        ]
       ],
       "shape": {
        "columns": 3,
        "rows": 8
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>value</th>\n",
       "      <th>LCL</th>\n",
       "      <th>UCL</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scale</th>\n",
       "      <th>coef</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">plain</th>\n",
       "      <th>b0</th>\n",
       "      <td>-197.434569</td>\n",
       "      <td>-402.432978</td>\n",
       "      <td>7.563840</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b1</th>\n",
       "      <td>248.490500</td>\n",
       "      <td>-14.566758</td>\n",
       "      <td>511.547758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b2</th>\n",
       "      <td>-98.599763</td>\n",
       "      <td>-205.469230</td>\n",
       "      <td>8.269703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b3</th>\n",
       "      <td>12.180353</td>\n",
       "      <td>-1.418593</td>\n",
       "      <td>25.779299</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">log</th>\n",
       "      <th>b0</th>\n",
       "      <td>-75.338681</td>\n",
       "      <td>-162.962520</td>\n",
       "      <td>12.285158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b1</th>\n",
       "      <td>287.517316</td>\n",
       "      <td>-51.624269</td>\n",
       "      <td>626.658902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b2</th>\n",
       "      <td>-339.655093</td>\n",
       "      <td>-748.869189</td>\n",
       "      <td>69.559003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>b3</th>\n",
       "      <td>124.514349</td>\n",
       "      <td>-29.183188</td>\n",
       "      <td>278.211886</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 value         LCL         UCL\n",
       "scale coef                                    \n",
       "plain b0   -197.434569 -402.432978    7.563840\n",
       "      b1    248.490500  -14.566758  511.547758\n",
       "      b2    -98.599763 -205.469230    8.269703\n",
       "      b3     12.180353   -1.418593   25.779299\n",
       "log   b0    -75.338681 -162.962520   12.285158\n",
       "      b1    287.517316  -51.624269  626.658902\n",
       "      b2   -339.655093 -748.869189   69.559003\n",
       "      b3    124.514349  -29.183188  278.211886"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# present table of results\n",
    "d = thetas.shape[-1]\n",
    "index = pd.MultiIndex.from_tuples([(lab, f\"b{i}\") for lab in scales for i in range(d)], names = ['scale', 'coef'])\n",
    "pd.DataFrame({\"value\": thetas.flatten(), \"LCL\": theta_CIs[:,0].flatten(), \"UCL\": theta_CIs[:,1].flatten()}, index = index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "b51346c478d5",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "scale",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "LLF",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "AIC",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "BIC",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "A",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "deviance",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "p-value(deviance_bootstrap)",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "deviance_bootstrap_samples",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "n",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "k",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "dof",
         "rawType": "int64",
         "type": "integer"
        }
       ],
       "ref": "6b22f7e6-a05d-41b0-8b8b-189ed817efb2",
       "rows": [
        [
         "plain",
         "-3.188295093857755",
         "14.376590187715511",
         "22.618362229901187",
         "0.9655172413793104",
         "6.37659018771551",
         "0.11188811188811189",
         "1000",
         "58",
         "4",
         "54"
        ],
        [
         "log",
         "-3.491409863193514",
         "14.982819726387028",
         "23.224591768572704",
         "0.9655172413793104",
         "6.982819726387028",
         "0.08191808191808192",
         "1000",
         "58",
         "4",
         "54"
        ]
       ],
       "shape": {
        "columns": 10,
        "rows": 2
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>LLF</th>\n",
       "      <th>AIC</th>\n",
       "      <th>BIC</th>\n",
       "      <th>A</th>\n",
       "      <th>deviance</th>\n",
       "      <th>p-value(deviance_bootstrap)</th>\n",
       "      <th>deviance_bootstrap_samples</th>\n",
       "      <th>n</th>\n",
       "      <th>k</th>\n",
       "      <th>dof</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scale</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>-3.188295</td>\n",
       "      <td>14.37659</td>\n",
       "      <td>22.618362</td>\n",
       "      <td>0.965517</td>\n",
       "      <td>6.37659</td>\n",
       "      <td>0.111888</td>\n",
       "      <td>1000</td>\n",
       "      <td>58</td>\n",
       "      <td>4</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>log</th>\n",
       "      <td>-3.491410</td>\n",
       "      <td>14.98282</td>\n",
       "      <td>23.224592</td>\n",
       "      <td>0.965517</td>\n",
       "      <td>6.98282</td>\n",
       "      <td>0.081918</td>\n",
       "      <td>1000</td>\n",
       "      <td>58</td>\n",
       "      <td>4</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            LLF       AIC        BIC         A  deviance  \\\n",
       "scale                                                      \n",
       "plain -3.188295  14.37659  22.618362  0.965517   6.37659   \n",
       "log   -3.491410  14.98282  23.224592  0.965517   6.98282   \n",
       "\n",
       "       p-value(deviance_bootstrap)  deviance_bootstrap_samples   n  k  dof  \n",
       "scale                                                                       \n",
       "plain                     0.111888                        1000  58  4   54  \n",
       "log                       0.081918                        1000  58  4   54  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# goodness of fit measures\n",
    "index = pd.Index(scales, name = 'scale')\n",
    "pd.DataFrame(\n",
    "    [logit.goodness_of_fit(x, y, theta, bootstrap_seed=random_seed)\n",
    "     for x, theta in zip(xs, thetas)],\n",
    "    index=index,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8029f7f521b1",
   "metadata": {},
   "source": [
    "### 3.2 Fitted risk curve and confidence bands\n",
    "\n",
    "**Coverage note.** The displayed limits are pointwise confidence intervals evaluated separately at each predictor value; they are not simultaneous confidence bands for the complete curve.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "141a540cbb5a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plotting results\n",
    "fig, axs = plt.subplots(ncols = len(scales), figsize = (6*len(scales), 5))\n",
    "\n",
    "quantile_methods = {\n",
    "    \"normal\": lambda x, p, theta, cov: logit.get_model_quantiles_normal(\n",
    "        x, p, theta, cov, seed=random_seed\n",
    "    ),\n",
    "    \"delta\": logit.get_model_quantiles_delta,\n",
    "}\n",
    "\n",
    "for ax, scale, x, theta, cov_theta in zip(axs, scales, xs, thetas, cov_thetas):\n",
    "\n",
    "    ax.set_title(f\"cond. prob. for AE: scale {scale}\")\n",
    "\n",
    "    if scale == \"plain\":\n",
    "        xlab = r\"$x = max_{visit} SUV(visit, p)$\"\n",
    "    else:\n",
    "        xlab = r\"$x = log(max_{visit} SUV(visit, p))$\"\n",
    "    \n",
    "    ax.set_xlabel(xlab)\n",
    "    ax.set_ylabel(\"P(AE|X = x)\")\n",
    "\n",
    "    # plotting data points\n",
    "    for i, group_label in enumerate([\"NC\", \"AE\"]):\n",
    "        ax.scatter(x[y == i], y[y == i], label=group_label)\n",
    "\n",
    "    # plotting fitted model \n",
    "    xp = np.linspace(min(x), max(x), 100)\n",
    "    yp = logit.model(xp, theta)\n",
    "\n",
    "    ax.plot(xp, yp, label = \"logistic reg\")\n",
    "\n",
    "    for ci_label, color in zip([\"normal\", \"delta\"], [\"red\", \"green\"]):\n",
    "        # get result of model quantiles at probs\n",
    "        # theta and cov_theta are obtained via MLE method\n",
    "        res = quantile_methods[ci_label](xp, probs, theta, cov_theta)\n",
    "\n",
    "        # make plot of quantiles\n",
    "        ax.fill_between(xp, *res, color=color, alpha=0.5, label=f\"CI:{ci_label}\")\n",
    "\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "fig.legend(handles, labels, bbox_to_anchor=(1.02, 0.5), loc='center right')\n",
    "\n",
    "fig.savefig(results_path / \"logit_cubic_nonmono_fit.pdf\", bbox_inches=\"tight\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "f57a346797fc",
   "metadata": {},
   "outputs": [],
   "source": [
    "cubic_unconstrained_results = {\n",
    "    \"model\": logit,\n",
    "    \"fit_results\": fit_results,\n",
    "    \"parameters\": thetas,\n",
    "    \"covariances\": cov_thetas,\n",
    "    \"parameter_ci_wald\": theta_CIs,\n",
    "}\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6b193fac8ac7",
   "metadata": {},
   "source": [
    "## 4. Cubic constrained (monotonic) logistic regression\n",
    "\n",
    "This model also uses cubic log-odds, but imposes the monotonicity condition\n",
    "\n",
    "$$\n",
    "\\eta'(t) \\ge 0\n",
    "$$\n",
    "\n",
    "over the fitted domain. It uses `degree=3`, `mono=True`, light L2 regularization, and differential-evolution fitting. In addition to the asymptotic analysis, this section compares confidence intervals obtained from normal parameter sampling, nonparametric bootstrap, stratified nonparametric bootstrap, and parametric bootstrap.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "2bf9dda5f4db",
   "metadata": {},
   "outputs": [],
   "source": [
    "results_path = results_root / \"cubic_monotonic\"\n",
    "results_path.mkdir(parents=True, exist_ok=True)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "50d789015352",
   "metadata": {},
   "source": [
    "### 4.1 Fit and numerical diagnostics\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "53ae4440fced",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "scale",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "theta",
         "rawType": "object",
         "type": "unknown"
        },
        {
         "name": "cost",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "success",
         "rawType": "bool",
         "type": "boolean"
        },
        {
         "name": "message",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "nit",
         "rawType": "int64",
         "type": "integer"
        }
       ],
       "ref": "c3049b51-6113-412c-9545-1d6c8a576605",
       "rows": [
        [
         "plain",
         "[-1.2399266532844526e+02 -4.3155748109311698e-05  5.5826774276550521e+00\n -1.2798214338792679e+01]",
         "4.782069566036849",
         "True",
         "CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH",
         "6"
        ],
        [
         "log",
         "[-4.0708789228957563e+01  1.2527203384049965e-08  1.5072426363862961e+01\n -1.2365102142378287e+01]",
         "4.60500749717172",
         "True",
         "CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*EPSMCH",
         "11"
        ]
       ],
       "shape": {
        "columns": 5,
        "rows": 2
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>theta</th>\n",
       "      <th>cost</th>\n",
       "      <th>success</th>\n",
       "      <th>message</th>\n",
       "      <th>nit</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scale</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>[-123.99266532844526, -4.31557481093117e-05, 5...</td>\n",
       "      <td>4.782070</td>\n",
       "      <td>True</td>\n",
       "      <td>CONVERGENCE: RELATIVE REDUCTION OF F &lt;= FACTR*...</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>log</th>\n",
       "      <td>[-40.70878922895756, 1.2527203384049965e-08, 1...</td>\n",
       "      <td>4.605007</td>\n",
       "      <td>True</td>\n",
       "      <td>CONVERGENCE: RELATIVE REDUCTION OF F &lt;= FACTR*...</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                   theta      cost  success  \\\n",
       "scale                                                                         \n",
       "plain  [-123.99266532844526, -4.31557481093117e-05, 5...  4.782070     True   \n",
       "log    [-40.70878922895756, 1.2527203384049965e-08, 1...  4.605007     True   \n",
       "\n",
       "                                                 message  nit  \n",
       "scale                                                          \n",
       "plain  CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*...    6  \n",
       "log    CONVERGENCE: RELATIVE REDUCTION OF F <= FACTR*...   11  "
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# constrained cubic logistic-regression model, light L2 regularization, monotonicity constraints\n",
    "logit = logistic.LogisticPolyRegression(degree=3, mono=True, lam=(0.0, 1e-6))\n",
    "\n",
    "# fits\n",
    "ress = [logit.fit(x, y, method=\"diff_evol\", seed=random_seed) for x in xs]\n",
    "if not all(result[\"success\"] for result in ress):\n",
    "    raise RuntimeError(\"At least one monotonic cubic logistic-regression fit failed.\")\n",
    "\n",
    "pd.DataFrame(ress, index = pd.Index(scales, name = 'scale'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "5cb58fb510b1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "theta:\n"
     ]
    },
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "scale",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "0",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "1",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "2",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "3",
         "rawType": "float64",
         "type": "float"
        }
       ],
       "ref": "82958ec5-0036-45f1-84be-822dbf1dae5c",
       "rows": [
        [
         "plain",
         "-123.99266532844526",
         "-4.31557481093117e-05",
         "5.582677427655052",
         "-12.798214338792679"
        ],
        [
         "log",
         "-40.70878922895756",
         "1.2527203384049965e-08",
         "15.072426363862961",
         "-12.365102142378287"
        ]
       ],
       "shape": {
        "columns": 4,
        "rows": 2
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scale</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-4.315575e-05</td>\n",
       "      <td>5.582677</td>\n",
       "      <td>-12.798214</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>log</th>\n",
       "      <td>-40.708789</td>\n",
       "      <td>1.252720e-08</td>\n",
       "      <td>15.072426</td>\n",
       "      <td>-12.365102</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                0             1          2          3\n",
       "scale                                                \n",
       "plain -123.992665 -4.315575e-05   5.582677 -12.798214\n",
       "log    -40.708789  1.252720e-08  15.072426 -12.365102"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "thetas = np.array([r[\"theta\"] for r in ress])\n",
    "\n",
    "print(\"theta:\")\n",
    "pd.DataFrame(thetas, index = pd.Index(scales, name = 'scale'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "5f9e98365ccd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "------------------------\n",
      "theta=[np.float64(-123.99266532844526), np.float64(-4.31557481093117e-05), np.float64(5.582677427655052), np.float64(-12.798214338792679)]\n",
      "beta=[np.float64(-123.99266532844526), np.float64(163.79429026354094), np.float64(-71.44830230346912), np.float64(10.38876242041641)]\n",
      "nllf = 4.7665004244040725, jac = [np.float64(0.0002489520620836405), np.float64(-2.4749811303377784e-05), np.float64(-4.583689623220699e-05), np.float64(-9.036528973715985e-06)]\n",
      "------------------------\n",
      "theta=[np.float64(-40.70878922895756), np.float64(1.2527203384049965e-08), np.float64(15.072426363862961), np.float64(-12.365102142378287)]\n",
      "beta=[np.float64(-40.70878922895756), np.float64(152.8957509914481), np.float64(-186.3720915226409), np.float64(75.72601216469042)]\n",
      "nllf = 4.6029702178637475, jac = [np.float64(8.141771471648074e-05), np.float64(3.0851046238366576e-09), np.float64(-3.014458160985464e-05), np.float64(2.4730816004896183e-05)]\n"
     ]
    }
   ],
   "source": [
    "for x, theta in zip(xs, thetas): \n",
    "    r = logit.get_nllf(x, y, theta, jac = True)\n",
    "    beta = logit.get_beta(theta)\n",
    "    print(\"------------------------\")\n",
    "    print(f\"theta={list(theta)}\")\n",
    "    print(f\"beta={list(beta)}\")\n",
    "    print(f\"nllf = {r[0]}, jac = {list(r[1])}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "89a497ed6bfb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cov_theta = array([[ 3.1859140165661383e+03, -2.9870601243316402e-04,\n",
      "        -7.6604317844440516e+01,  1.6819875706593353e+02],\n",
      "       [-2.9870601243316402e-04,  6.9641901741636447e-10,\n",
      "         1.0490946865335382e-05, -1.8139421543644511e-05],\n",
      "       [-7.6604317844440516e+01,  1.0490946865335382e-05,\n",
      "         1.8957780916257068e+00, -4.0834134155788258e+00],\n",
      "       [ 1.6819875706593353e+02, -1.8139421543644511e-05,\n",
      "        -4.0834134155788258e+00,  8.9092792565209358e+00]])\n",
      "cov_theta = array([[ 4.1490658150157788e+02,  3.8296900815642823e-08,\n",
      "        -8.4427289013825401e+01,  6.4969656267749656e+01],\n",
      "       [ 3.8296900815642823e-08,  8.0429626556602896e-17,\n",
      "        -1.3744208941948679e-08,  7.4712782907274201e-09],\n",
      "       [-8.4427289013825401e+01, -1.3744208941948679e-08,\n",
      "         1.8371717255735291e+01, -1.3529611929618980e+01],\n",
      "       [ 6.4969656267749656e+01,  7.4712782907274201e-09,\n",
      "        -1.3529611929618980e+01,  1.0261562792563552e+01]])\n"
     ]
    }
   ],
   "source": [
    "# asymptotic covariance matrix of parameters\n",
    "cov_thetas = np.array([logit.get_cov(x, y, theta) for x, theta in zip(xs, thetas)])\n",
    "\n",
    "for cov_theta in cov_thetas: print(f\"{cov_theta = }\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "16a0f1a3981a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[4.6377101945380813e-10 5.7407317504966392e-04 8.2374363426882577e-02\n",
      " 3.1966361254779154e+03]\n",
      "[1.6653345369190556e-15 7.2506809467669259e-03 1.2058261277701048e+00\n",
      " 4.4232678474115988e+02]\n"
     ]
    }
   ],
   "source": [
    "# check eigenvalues of the symmetric covariance matrices\n",
    "for cov_theta in cov_thetas:\n",
    "    print(np.linalg.eigvalsh(cov_theta))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c2ee2e94167d",
   "metadata": {},
   "source": [
    "### 4.2 Wald theta-parameter intervals and goodness of fit\n",
    "\n",
    "This subsection reports uncertainty for the optimization parameters $\\theta$, not for the polynomial coefficients $\\beta$. In the monotonic cubic model the two parameterizations differ.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "ed2a1bcb6679",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "theta_CI = array([[-2.3462067941579588e+02, -9.4878711842582338e-05,\n",
      "         2.8840567666220402e+00, -1.8648396334764222e+01],\n",
      "       [-1.3364651241094663e+01,  8.5672156239589289e-06,\n",
      "         8.2812980886880645e+00, -6.9480323428211390e+00]])\n",
      "theta_CI = array([[-8.0631796972278750e+01, -5.0502565684288599e-09,\n",
      "         6.6715814478236926e+00, -1.8643586904134096e+01],\n",
      "       [-7.8578148563639161e-01,  3.0104663336528783e-08,\n",
      "         2.3473271279902228e+01, -6.0866173806224779e+00]])\n"
     ]
    }
   ],
   "source": [
    "# CI of params (assuming asymptotic distr of parameters) : Wald approximation\n",
    "theta_CIs = np.array([\n",
    "    logit.get_theta_quantiles_normal(probs, theta, cov_theta)\n",
    "    for theta, cov_theta in zip(thetas, cov_thetas)\n",
    "])\n",
    "\n",
    "for theta_CI in theta_CIs: print(f\"{theta_CI = }\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "e74159b26dc8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "index",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "scale",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "coef",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "value",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "LCL",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "UCL",
         "rawType": "float64",
         "type": "float"
        }
       ],
       "ref": "ac60d1db-ae27-4785-a209-cbf235b1ad44",
       "rows": [
        [
         "0",
         "plain",
         "p0",
         "-123.99266532844526",
         "-234.62067941579588",
         "-13.364651241094663"
        ],
        [
         "1",
         "plain",
         "p1",
         "-4.31557481093117e-05",
         "-9.487871184258234e-05",
         "8.567215623958929e-06"
        ],
        [
         "2",
         "plain",
         "p2",
         "5.582677427655052",
         "2.8840567666220402",
         "8.281298088688064"
        ],
        [
         "3",
         "plain",
         "p3",
         "-12.798214338792679",
         "-18.64839633476422",
         "-6.948032342821139"
        ],
        [
         "4",
         "log",
         "p0",
         "-40.70878922895756",
         "-80.63179697227875",
         "-0.7857814856363916"
        ],
        [
         "5",
         "log",
         "p1",
         "1.2527203384049965e-08",
         "-5.05025656842886e-09",
         "3.010466333652878e-08"
        ],
        [
         "6",
         "log",
         "p2",
         "15.072426363862961",
         "6.671581447823693",
         "23.473271279902228"
        ],
        [
         "7",
         "log",
         "p3",
         "-12.365102142378287",
         "-18.643586904134096",
         "-6.086617380622478"
        ]
       ],
       "shape": {
        "columns": 5,
        "rows": 8
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>scale</th>\n",
       "      <th>coef</th>\n",
       "      <th>value</th>\n",
       "      <th>LCL</th>\n",
       "      <th>UCL</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>plain</td>\n",
       "      <td>p0</td>\n",
       "      <td>-1.239927e+02</td>\n",
       "      <td>-2.346207e+02</td>\n",
       "      <td>-1.336465e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>plain</td>\n",
       "      <td>p1</td>\n",
       "      <td>-4.315575e-05</td>\n",
       "      <td>-9.487871e-05</td>\n",
       "      <td>8.567216e-06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>plain</td>\n",
       "      <td>p2</td>\n",
       "      <td>5.582677e+00</td>\n",
       "      <td>2.884057e+00</td>\n",
       "      <td>8.281298e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>plain</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.279821e+01</td>\n",
       "      <td>-1.864840e+01</td>\n",
       "      <td>-6.948032e+00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>log</td>\n",
       "      <td>p0</td>\n",
       "      <td>-4.070879e+01</td>\n",
       "      <td>-8.063180e+01</td>\n",
       "      <td>-7.857815e-01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>log</td>\n",
       "      <td>p1</td>\n",
       "      <td>1.252720e-08</td>\n",
       "      <td>-5.050257e-09</td>\n",
       "      <td>3.010466e-08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>log</td>\n",
       "      <td>p2</td>\n",
       "      <td>1.507243e+01</td>\n",
       "      <td>6.671581e+00</td>\n",
       "      <td>2.347327e+01</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>log</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.236510e+01</td>\n",
       "      <td>-1.864359e+01</td>\n",
       "      <td>-6.086617e+00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   scale coef         value           LCL           UCL\n",
       "0  plain   p0 -1.239927e+02 -2.346207e+02 -1.336465e+01\n",
       "1  plain   p1 -4.315575e-05 -9.487871e-05  8.567216e-06\n",
       "2  plain   p2  5.582677e+00  2.884057e+00  8.281298e+00\n",
       "3  plain   p3 -1.279821e+01 -1.864840e+01 -6.948032e+00\n",
       "4    log   p0 -4.070879e+01 -8.063180e+01 -7.857815e-01\n",
       "5    log   p1  1.252720e-08 -5.050257e-09  3.010466e-08\n",
       "6    log   p2  1.507243e+01  6.671581e+00  2.347327e+01\n",
       "7    log   p3 -1.236510e+01 -1.864359e+01 -6.086617e+00"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# nr of parameters\n",
    "d = thetas.shape[-1]\n",
    "\n",
    "# present table of results\n",
    "df_theta_CI_wald = pd.DataFrame({\n",
    "    \"scale\" :[lab for lab in scales for _ in range(d)],\n",
    "    \"coef\" : [ f\"p{i}\" for _ in scales for i in range(d)],\n",
    "    \"value\" : thetas.flatten(), \n",
    "    \"LCL\" : theta_CIs[:,0].flatten(), \n",
    "    \"UCL\": theta_CIs[:,1].flatten()\n",
    "})\n",
    "\n",
    "df_theta_CI_wald"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "2f192e86182d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "scale",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "LLF",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "AIC",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "BIC",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "A",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "deviance",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "p-value(deviance_bootstrap)",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "deviance_bootstrap_samples",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "n",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "k",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "dof",
         "rawType": "int64",
         "type": "integer"
        }
       ],
       "ref": "b4fbadb9-a46a-45c1-bfdd-aaecefe4ab96",
       "rows": [
        [
         "plain",
         "-4.7665004244040725",
         "17.533000848808143",
         "25.774772890993823",
         "0.9655172413793104",
         "9.533000848808145",
         "0.20679320679320679",
         "1000",
         "58",
         "4",
         "54"
        ],
        [
         "log",
         "-4.6029702178637475",
         "17.205940435727495",
         "25.44771247791317",
         "0.9482758620689655",
         "9.205940435727495",
         "0.19080919080919082",
         "1000",
         "58",
         "4",
         "54"
        ]
       ],
       "shape": {
        "columns": 10,
        "rows": 2
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>LLF</th>\n",
       "      <th>AIC</th>\n",
       "      <th>BIC</th>\n",
       "      <th>A</th>\n",
       "      <th>deviance</th>\n",
       "      <th>p-value(deviance_bootstrap)</th>\n",
       "      <th>deviance_bootstrap_samples</th>\n",
       "      <th>n</th>\n",
       "      <th>k</th>\n",
       "      <th>dof</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>scale</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>plain</th>\n",
       "      <td>-4.76650</td>\n",
       "      <td>17.533001</td>\n",
       "      <td>25.774773</td>\n",
       "      <td>0.965517</td>\n",
       "      <td>9.533001</td>\n",
       "      <td>0.206793</td>\n",
       "      <td>1000</td>\n",
       "      <td>58</td>\n",
       "      <td>4</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>log</th>\n",
       "      <td>-4.60297</td>\n",
       "      <td>17.205940</td>\n",
       "      <td>25.447712</td>\n",
       "      <td>0.948276</td>\n",
       "      <td>9.205940</td>\n",
       "      <td>0.190809</td>\n",
       "      <td>1000</td>\n",
       "      <td>58</td>\n",
       "      <td>4</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           LLF        AIC        BIC         A  deviance  \\\n",
       "scale                                                      \n",
       "plain -4.76650  17.533001  25.774773  0.965517  9.533001   \n",
       "log   -4.60297  17.205940  25.447712  0.948276  9.205940   \n",
       "\n",
       "       p-value(deviance_bootstrap)  deviance_bootstrap_samples   n  k  dof  \n",
       "scale                                                                       \n",
       "plain                     0.206793                        1000  58  4   54  \n",
       "log                       0.190809                        1000  58  4   54  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# goodness of fit measures\n",
    "pd.DataFrame(\n",
    "    [logit.goodness_of_fit(x, y, theta, bootstrap_seed=random_seed)\n",
    "     for x, theta in zip(xs, thetas)], \n",
    "    index = pd.Index(scales, name = 'scale')\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ad4c94ba24a",
   "metadata": {},
   "source": [
    "### 4.3 Fitted risk curve and asymptotic confidence bands\n",
    "\n",
    "**Coverage note.** The displayed limits are pointwise confidence intervals evaluated separately at each predictor value; they are not simultaneous confidence bands for the complete curve.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "91ed23f07d87",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plotting results\n",
    "fig, axs = plt.subplots(ncols = len(scales), figsize = (6*len(scales), 5))\n",
    "\n",
    "quantile_methods = {\n",
    "    \"normal\": lambda x, p, theta, cov: logit.get_model_quantiles_normal(\n",
    "        x, p, theta, cov, seed=random_seed\n",
    "    ),\n",
    "    \"delta\": logit.get_model_quantiles_delta,\n",
    "}\n",
    "\n",
    "for ax, scale, x, theta, cov_theta in zip(axs, scales, xs, thetas, cov_thetas):\n",
    "\n",
    "    ax.set_title(f\"cond. prob. for AE: scale {scale}\")\n",
    "\n",
    "    if scale == \"plain\":\n",
    "        xlab = r\"$x = max_{visit} SUV(visit, p)$\"\n",
    "    else:\n",
    "        xlab = r\"$x = log(max_{visit} SUV(visit, p))$\"\n",
    "    \n",
    "    ax.set_xlabel(xlab)\n",
    "    ax.set_ylabel(\"P(AE|X = x)\")\n",
    "\n",
    "    # plotting data points\n",
    "    for i, group_label in enumerate([\"NC\", \"AE\"]):\n",
    "        ax.scatter(x[y == i], y[y == i], label=group_label)\n",
    "\n",
    "    # plotting fitted model \n",
    "    dx = max(x) - min(x)\n",
    "    xp = np.linspace(min(x) - 0.1 * dx, max(x) + 0.1 * dx, 100)\n",
    "    yp = logit.model(xp, theta)\n",
    "\n",
    "    ax.plot(xp, yp, label = \"logistic reg\")\n",
    "\n",
    "    for ci_label, color in zip([\"normal\", \"delta\"], [\"red\", \"green\"]):\n",
    "        # get result of model quantiles at probs\n",
    "        # theta and cov_theta are obtained via MLE method\n",
    "        res = quantile_methods[ci_label](xp, probs, theta, cov_theta)\n",
    "\n",
    "        # make plot of quantiles\n",
    "        ax.fill_between(xp, *res, color=color, alpha=0.5, label=f\"CI:{ci_label}\")\n",
    "\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "fig.legend(handles, labels, bbox_to_anchor=(1.02, 0.5), loc='center right')\n",
    "\n",
    "fig.savefig(results_path / \"logit_mono_asymptotic_fit.pdf\", bbox_inches=\"tight\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "32da73e5cac3",
   "metadata": {},
   "source": [
    "### 4.4 Model confidence-interval comparison\n",
    "\n",
    "**Coverage note.** All analytic and bootstrap envelopes compared below are pointwise in the predictor. Their nominal coverage applies at a fixed predictor value, not simultaneously over the complete curve.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "cfe79b472ce4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model CI:scale:plain,method:normal\n",
      "model CI:scale:plain,method:delta\n",
      "model CI:scale:plain,boots-method:normal\n",
      "model CI:scale:plain,boots-method:nonparam_boots\n",
      "model CI:scale:plain,boots-method:nonparam_stratified_boots\n",
      "model CI:scale:plain,boots-method:parametric_boots\n",
      "model CI:scale:log,method:normal\n",
      "model CI:scale:log,method:delta\n",
      "model CI:scale:log,boots-method:normal\n",
      "model CI:scale:log,boots-method:nonparam_boots\n",
      "model CI:scale:log,boots-method:nonparam_stratified_boots\n",
      "model CI:scale:log,boots-method:parametric_boots\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plotting results\n",
    "fig, axs = plt.subplots(ncols = len(scales), figsize = (6*len(scales), 5))\n",
    "\n",
    "boots_theta_results = {}\n",
    "quantile_methods = {\n",
    "    \"normal\": lambda x, p, theta, cov: logit.get_model_quantiles_normal(\n",
    "        x, p, theta, cov, seed=random_seed\n",
    "    ),\n",
    "    \"delta\": logit.get_model_quantiles_delta,\n",
    "}\n",
    "parameter_generators = {\n",
    "    \"normal\": logit.get_normal_theta,\n",
    "    \"nonparam_boots\": logit.get_nonparam_boots_theta,\n",
    "    \"nonparam_stratified_boots\": logit.get_nonparam_stratified_boots_theta,\n",
    "    \"parametric_boots\": logit.get_parametric_boots_theta,\n",
    "}\n",
    "n_parameter_samples = 10_000\n",
    "\n",
    "for ax, scale, x, theta, cov_theta in zip(axs, scales, xs, thetas, cov_thetas):\n",
    "\n",
    "    ax.set_title(f\"cond. prob. for AE: scale {scale}\")\n",
    "    \n",
    "    if scale == \"plain\":\n",
    "        xlab = r\"$x = max_{visit} SUV(visit, p)$\"\n",
    "    else:\n",
    "        xlab = r\"$x = log(max_{visit} SUV(visit, p))$\"\n",
    "    \n",
    "    ax.set_xlabel(xlab)\n",
    "    ax.set_ylabel(\"P(AE|X = x)\")\n",
    "\n",
    "    # plotting data points\n",
    "    for i, group_label in enumerate([\"NC\", \"AE\"]):\n",
    "        ax.scatter(x[y == i], y[y == i], label=group_label)\n",
    "\n",
    "    # plotting fitted model\n",
    "    dx = max(x) - min(x)\n",
    "    xp = np.linspace(min(x) - 0.1 * dx, max(x) + 0.1 * dx, 100)\n",
    "    yp = logit.model(xp, theta)\n",
    "\n",
    "    ax.plot(xp, yp, label = \"logistic reg\")\n",
    "\n",
    "    for method, color in zip([\"normal\", \"delta\"], [\"red\", \"green\"]):\n",
    "        \n",
    "        print(f\"model CI:scale:{scale},method:{method}\")\n",
    "\n",
    "        # get result of model quantiles at probs\n",
    "        # theta and cov_theta are obtained via MLE method\n",
    "        res = quantile_methods[method](xp, probs, theta, cov_theta)\n",
    "\n",
    "        # make plot of quantiles\n",
    "        ax.fill_between(xp, *res, color=color, alpha=0.5, label=f\"CI:{method}\")\n",
    "\n",
    "    # plot quantiles of models at sampled or bootstrapped parameters\n",
    "    for method, color in zip(\n",
    "        [\"normal\", \"nonparam_boots\", \"nonparam_stratified_boots\", \"parametric_boots\"],\n",
    "        [\"orange\", \"pink\", \"purple\", \"cyan\"]):\n",
    "    \n",
    "        print(f\"model CI:scale:{scale},boots-method:{method}\")\n",
    "\n",
    "        # generate sampled or bootstrapped parameters\n",
    "        btheta = parameter_generators[method](\n",
    "            x, y, n_parameter_samples, seed=random_seed\n",
    "        )\n",
    "    \n",
    "        boots_theta_results[(scale, method)] = btheta\n",
    "\n",
    "        # quantiles\n",
    "        quant = np.quantile([logit.model(xp, p) for p in btheta], probs, axis = 0)\n",
    "\n",
    "        # make plot of quantiles\n",
    "        ax.fill_between(xp, *quant, color=color, alpha=0.5, label=f\"CI:boots, {method}\")\n",
    "\n",
    "handles, labels = ax.get_legend_handles_labels()\n",
    "fig.legend(handles, labels, bbox_to_anchor=(1.17, 0.5), loc='center right')\n",
    "\n",
    "fig.savefig(results_path / \"logit_mono_ci_comparison.pdf\", bbox_inches=\"tight\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "121501281f2e",
   "metadata": {},
   "source": [
    "### 4.5 Theta-parameter confidence-interval comparison\n",
    "\n",
    "The comparison below is the **previous $\\theta$-parameter analysis**: the Wald limits and all sampled/bootstrap quantiles are calculated component by component for $\\theta$. The existing labels `p0`--`p3` therefore denote $\\theta_0$--$\\theta_3$, not the polynomial coefficients $\\beta_0$--$\\beta_3$. Because the monotonic parameterization is non-unique, these component-wise $\\theta$ intervals are useful mainly as numerical diagnostics and are not reliably interpretable as coefficient intervals.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "2ed7732fc9ea",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "theta CI:scale:['plain', 'log'],method:Wald/delta\n",
      "model CI:scale:plain,boots-method:normal\n",
      "model CI:scale:plain,boots-method:nonparam_boots\n",
      "model CI:scale:plain,boots-method:nonparam_stratified_boots\n",
      "model CI:scale:plain,boots-method:parametric_boots\n",
      "model CI:scale:log,boots-method:normal\n",
      "model CI:scale:log,boots-method:nonparam_boots\n",
      "model CI:scale:log,boots-method:nonparam_stratified_boots\n",
      "model CI:scale:log,boots-method:parametric_boots\n"
     ]
    },
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "index",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "scale",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "coef",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "value",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "LCL",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "UCL",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "method",
         "rawType": "object",
         "type": "string"
        }
       ],
       "ref": "17c32172-7132-45ba-974f-2559ab2ef80a",
       "rows": [
        [
         "0",
         "plain",
         "p0",
         "-123.99266532844526",
         "-234.62067941579588",
         "-13.364651241094663",
         "Wald/delta"
        ],
        [
         "1",
         "plain",
         "p1",
         "-4.31557481093117e-05",
         "-9.487871184258234e-05",
         "8.567215623958929e-06",
         "Wald/delta"
        ],
        [
         "2",
         "plain",
         "p2",
         "5.582677427655052",
         "2.8840567666220402",
         "8.281298088688064",
         "Wald/delta"
        ],
        [
         "3",
         "plain",
         "p3",
         "-12.798214338792679",
         "-18.64839633476422",
         "-6.948032342821139",
         "Wald/delta"
        ],
        [
         "4",
         "log",
         "p0",
         "-40.70878922895756",
         "-80.63179697227875",
         "-0.7857814856363916",
         "Wald/delta"
        ],
        [
         "5",
         "log",
         "p1",
         "1.2527203384049965e-08",
         "-5.05025656842886e-09",
         "3.010466333652878e-08",
         "Wald/delta"
        ],
        [
         "6",
         "log",
         "p2",
         "15.072426363862961",
         "6.671581447823693",
         "23.473271279902228",
         "Wald/delta"
        ],
        [
         "7",
         "log",
         "p3",
         "-12.365102142378287",
         "-18.643586904134096",
         "-6.086617380622478",
         "Wald/delta"
        ],
        [
         "8",
         "plain",
         "p0",
         "-123.99266532844526",
         "-233.94943123214563",
         "-9.562456807989498",
         "normal"
        ],
        [
         "9",
         "plain",
         "p1",
         "-4.31557481093117e-05",
         "-9.595711131465509e-05",
         "7.713982008034643e-06",
         "normal"
        ],
        [
         "10",
         "plain",
         "p2",
         "5.582677427655052",
         "2.7944366226901765",
         "8.270209121704738",
         "normal"
        ],
        [
         "11",
         "plain",
         "p3",
         "-12.798214338792679",
         "-18.63648977078935",
         "-6.719472346319873",
         "normal"
        ],
        [
         "12",
         "plain",
         "p0",
         "-123.99266532844526",
         "-662.077284728864",
         "-20.9316993950732",
         "nonparam_boots"
        ],
        [
         "13",
         "plain",
         "p1",
         "-4.31557481093117e-05",
         "-2.42017933708336",
         "0.0003871658766217858",
         "nonparam_boots"
        ],
        [
         "14",
         "plain",
         "p2",
         "5.582677427655052",
         "-11.027819990575658",
         "17.817659569531223",
         "nonparam_boots"
        ],
        [
         "15",
         "plain",
         "p3",
         "-12.798214338792679",
         "-32.8388006577751",
         "5.78917152016678",
         "nonparam_boots"
        ],
        [
         "16",
         "plain",
         "p0",
         "-123.99266532844526",
         "-647.740989994033",
         "-21.412504234536605",
         "nonparam_stratified_boots"
        ],
        [
         "17",
         "plain",
         "p1",
         "-4.31557481093117e-05",
         "-2.530518347267342",
         "0.21072214516873888",
         "nonparam_stratified_boots"
        ],
        [
         "18",
         "plain",
         "p2",
         "5.582677427655052",
         "-11.034868065816198",
         "17.537220601664792",
         "nonparam_stratified_boots"
        ],
        [
         "19",
         "plain",
         "p3",
         "-12.798214338792679",
         "-32.43663897803941",
         "5.801689872262745",
         "nonparam_stratified_boots"
        ],
        [
         "20",
         "plain",
         "p0",
         "-123.99266532844526",
         "-519.9185959807692",
         "-14.267927545473208",
         "parametric_boots"
        ],
        [
         "21",
         "plain",
         "p1",
         "-4.31557481093117e-05",
         "-2.558558838295493",
         "0.017587628573496706",
         "parametric_boots"
        ],
        [
         "22",
         "plain",
         "p2",
         "5.582677427655052",
         "-17.294438383114326",
         "14.72733658434196",
         "parametric_boots"
        ],
        [
         "23",
         "plain",
         "p3",
         "-12.798214338792679",
         "-28.083488712335097",
         "26.4111328579993",
         "parametric_boots"
        ],
        [
         "24",
         "log",
         "p0",
         "-40.70878922895756",
         "-80.51482391483258",
         "0.5768096898392844",
         "normal"
        ],
        [
         "25",
         "log",
         "p1",
         "1.2527203384049965e-08",
         "-4.475512960882313e-09",
         "2.942177378758347e-08",
         "normal"
        ],
        [
         "26",
         "log",
         "p2",
         "15.072426363862961",
         "6.428671163878268",
         "23.468982296502062",
         "normal"
        ],
        [
         "27",
         "log",
         "p3",
         "-12.365102142378287",
         "-18.600564651751515",
         "-5.869816698776138",
         "normal"
        ],
        [
         "28",
         "log",
         "p0",
         "-40.70878922895756",
         "-362.3418986772154",
         "-14.457646430958098",
         "nonparam_boots"
        ],
        [
         "29",
         "log",
         "p1",
         "1.2527203384049965e-08",
         "-1.2949692152748325e-05",
         "1.3701590772882096e-05",
         "nonparam_boots"
        ],
        [
         "30",
         "log",
         "p2",
         "15.072426363862961",
         "-35.74117319597362",
         "78.60083374376462",
         "nonparam_boots"
        ],
        [
         "31",
         "log",
         "p3",
         "-12.365102142378287",
         "-44.06299924037829",
         "2.2799625576498834",
         "nonparam_boots"
        ],
        [
         "32",
         "log",
         "p0",
         "-40.70878922895756",
         "-369.35112799570214",
         "-14.205808684448709",
         "nonparam_stratified_boots"
        ],
        [
         "33",
         "log",
         "p1",
         "1.2527203384049965e-08",
         "-1.1407057756070862e-05",
         "1.2099347078543717e-05",
         "nonparam_stratified_boots"
        ],
        [
         "34",
         "log",
         "p2",
         "15.072426363862961",
         "-35.75099430893984",
         "79.27185653189909",
         "nonparam_stratified_boots"
        ],
        [
         "35",
         "log",
         "p3",
         "-12.365102142378287",
         "-44.49996309035288",
         "2.2802825280843164",
         "nonparam_stratified_boots"
        ],
        [
         "36",
         "log",
         "p0",
         "-40.70878922895756",
         "-310.97179686469656",
         "-8.945695488818114",
         "parametric_boots"
        ],
        [
         "37",
         "log",
         "p1",
         "1.2527203384049965e-08",
         "-0.00022031918968224803",
         "0.00024100793866809866",
         "parametric_boots"
        ],
        [
         "38",
         "log",
         "p2",
         "15.072426363862961",
         "-74.06005091397371",
         "48.61999601158131",
         "parametric_boots"
        ],
        [
         "39",
         "log",
         "p3",
         "-12.365102142378287",
         "-34.97653145139742",
         "17.689024661963625",
         "parametric_boots"
        ]
       ],
       "shape": {
        "columns": 6,
        "rows": 40
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>scale</th>\n",
       "      <th>coef</th>\n",
       "      <th>value</th>\n",
       "      <th>LCL</th>\n",
       "      <th>UCL</th>\n",
       "      <th>method</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>plain</td>\n",
       "      <td>p0</td>\n",
       "      <td>-1.239927e+02</td>\n",
       "      <td>-2.346207e+02</td>\n",
       "      <td>-1.336465e+01</td>\n",
       "      <td>Wald/delta</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>plain</td>\n",
       "      <td>p1</td>\n",
       "      <td>-4.315575e-05</td>\n",
       "      <td>-9.487871e-05</td>\n",
       "      <td>8.567216e-06</td>\n",
       "      <td>Wald/delta</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>plain</td>\n",
       "      <td>p2</td>\n",
       "      <td>5.582677e+00</td>\n",
       "      <td>2.884057e+00</td>\n",
       "      <td>8.281298e+00</td>\n",
       "      <td>Wald/delta</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>plain</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.279821e+01</td>\n",
       "      <td>-1.864840e+01</td>\n",
       "      <td>-6.948032e+00</td>\n",
       "      <td>Wald/delta</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>log</td>\n",
       "      <td>p0</td>\n",
       "      <td>-4.070879e+01</td>\n",
       "      <td>-8.063180e+01</td>\n",
       "      <td>-7.857815e-01</td>\n",
       "      <td>Wald/delta</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>log</td>\n",
       "      <td>p1</td>\n",
       "      <td>1.252720e-08</td>\n",
       "      <td>-5.050257e-09</td>\n",
       "      <td>3.010466e-08</td>\n",
       "      <td>Wald/delta</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>log</td>\n",
       "      <td>p2</td>\n",
       "      <td>1.507243e+01</td>\n",
       "      <td>6.671581e+00</td>\n",
       "      <td>2.347327e+01</td>\n",
       "      <td>Wald/delta</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>log</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.236510e+01</td>\n",
       "      <td>-1.864359e+01</td>\n",
       "      <td>-6.086617e+00</td>\n",
       "      <td>Wald/delta</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>plain</td>\n",
       "      <td>p0</td>\n",
       "      <td>-1.239927e+02</td>\n",
       "      <td>-2.339494e+02</td>\n",
       "      <td>-9.562457e+00</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>plain</td>\n",
       "      <td>p1</td>\n",
       "      <td>-4.315575e-05</td>\n",
       "      <td>-9.595711e-05</td>\n",
       "      <td>7.713982e-06</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>plain</td>\n",
       "      <td>p2</td>\n",
       "      <td>5.582677e+00</td>\n",
       "      <td>2.794437e+00</td>\n",
       "      <td>8.270209e+00</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>plain</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.279821e+01</td>\n",
       "      <td>-1.863649e+01</td>\n",
       "      <td>-6.719472e+00</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>plain</td>\n",
       "      <td>p0</td>\n",
       "      <td>-1.239927e+02</td>\n",
       "      <td>-6.620773e+02</td>\n",
       "      <td>-2.093170e+01</td>\n",
       "      <td>nonparam_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>plain</td>\n",
       "      <td>p1</td>\n",
       "      <td>-4.315575e-05</td>\n",
       "      <td>-2.420179e+00</td>\n",
       "      <td>3.871659e-04</td>\n",
       "      <td>nonparam_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>plain</td>\n",
       "      <td>p2</td>\n",
       "      <td>5.582677e+00</td>\n",
       "      <td>-1.102782e+01</td>\n",
       "      <td>1.781766e+01</td>\n",
       "      <td>nonparam_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>plain</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.279821e+01</td>\n",
       "      <td>-3.283880e+01</td>\n",
       "      <td>5.789172e+00</td>\n",
       "      <td>nonparam_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>plain</td>\n",
       "      <td>p0</td>\n",
       "      <td>-1.239927e+02</td>\n",
       "      <td>-6.477410e+02</td>\n",
       "      <td>-2.141250e+01</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>plain</td>\n",
       "      <td>p1</td>\n",
       "      <td>-4.315575e-05</td>\n",
       "      <td>-2.530518e+00</td>\n",
       "      <td>2.107221e-01</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>plain</td>\n",
       "      <td>p2</td>\n",
       "      <td>5.582677e+00</td>\n",
       "      <td>-1.103487e+01</td>\n",
       "      <td>1.753722e+01</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>plain</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.279821e+01</td>\n",
       "      <td>-3.243664e+01</td>\n",
       "      <td>5.801690e+00</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>plain</td>\n",
       "      <td>p0</td>\n",
       "      <td>-1.239927e+02</td>\n",
       "      <td>-5.199186e+02</td>\n",
       "      <td>-1.426793e+01</td>\n",
       "      <td>parametric_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>plain</td>\n",
       "      <td>p1</td>\n",
       "      <td>-4.315575e-05</td>\n",
       "      <td>-2.558559e+00</td>\n",
       "      <td>1.758763e-02</td>\n",
       "      <td>parametric_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>plain</td>\n",
       "      <td>p2</td>\n",
       "      <td>5.582677e+00</td>\n",
       "      <td>-1.729444e+01</td>\n",
       "      <td>1.472734e+01</td>\n",
       "      <td>parametric_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>plain</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.279821e+01</td>\n",
       "      <td>-2.808349e+01</td>\n",
       "      <td>2.641113e+01</td>\n",
       "      <td>parametric_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>log</td>\n",
       "      <td>p0</td>\n",
       "      <td>-4.070879e+01</td>\n",
       "      <td>-8.051482e+01</td>\n",
       "      <td>5.768097e-01</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>log</td>\n",
       "      <td>p1</td>\n",
       "      <td>1.252720e-08</td>\n",
       "      <td>-4.475513e-09</td>\n",
       "      <td>2.942177e-08</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>log</td>\n",
       "      <td>p2</td>\n",
       "      <td>1.507243e+01</td>\n",
       "      <td>6.428671e+00</td>\n",
       "      <td>2.346898e+01</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>log</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.236510e+01</td>\n",
       "      <td>-1.860056e+01</td>\n",
       "      <td>-5.869817e+00</td>\n",
       "      <td>normal</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>log</td>\n",
       "      <td>p0</td>\n",
       "      <td>-4.070879e+01</td>\n",
       "      <td>-3.623419e+02</td>\n",
       "      <td>-1.445765e+01</td>\n",
       "      <td>nonparam_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>log</td>\n",
       "      <td>p1</td>\n",
       "      <td>1.252720e-08</td>\n",
       "      <td>-1.294969e-05</td>\n",
       "      <td>1.370159e-05</td>\n",
       "      <td>nonparam_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>log</td>\n",
       "      <td>p2</td>\n",
       "      <td>1.507243e+01</td>\n",
       "      <td>-3.574117e+01</td>\n",
       "      <td>7.860083e+01</td>\n",
       "      <td>nonparam_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>log</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.236510e+01</td>\n",
       "      <td>-4.406300e+01</td>\n",
       "      <td>2.279963e+00</td>\n",
       "      <td>nonparam_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>log</td>\n",
       "      <td>p0</td>\n",
       "      <td>-4.070879e+01</td>\n",
       "      <td>-3.693511e+02</td>\n",
       "      <td>-1.420581e+01</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>log</td>\n",
       "      <td>p1</td>\n",
       "      <td>1.252720e-08</td>\n",
       "      <td>-1.140706e-05</td>\n",
       "      <td>1.209935e-05</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>log</td>\n",
       "      <td>p2</td>\n",
       "      <td>1.507243e+01</td>\n",
       "      <td>-3.575099e+01</td>\n",
       "      <td>7.927186e+01</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>log</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.236510e+01</td>\n",
       "      <td>-4.449996e+01</td>\n",
       "      <td>2.280283e+00</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>log</td>\n",
       "      <td>p0</td>\n",
       "      <td>-4.070879e+01</td>\n",
       "      <td>-3.109718e+02</td>\n",
       "      <td>-8.945695e+00</td>\n",
       "      <td>parametric_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>log</td>\n",
       "      <td>p1</td>\n",
       "      <td>1.252720e-08</td>\n",
       "      <td>-2.203192e-04</td>\n",
       "      <td>2.410079e-04</td>\n",
       "      <td>parametric_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>log</td>\n",
       "      <td>p2</td>\n",
       "      <td>1.507243e+01</td>\n",
       "      <td>-7.406005e+01</td>\n",
       "      <td>4.862000e+01</td>\n",
       "      <td>parametric_boots</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>log</td>\n",
       "      <td>p3</td>\n",
       "      <td>-1.236510e+01</td>\n",
       "      <td>-3.497653e+01</td>\n",
       "      <td>1.768902e+01</td>\n",
       "      <td>parametric_boots</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    scale coef         value           LCL           UCL  \\\n",
       "0   plain   p0 -1.239927e+02 -2.346207e+02 -1.336465e+01   \n",
       "1   plain   p1 -4.315575e-05 -9.487871e-05  8.567216e-06   \n",
       "2   plain   p2  5.582677e+00  2.884057e+00  8.281298e+00   \n",
       "3   plain   p3 -1.279821e+01 -1.864840e+01 -6.948032e+00   \n",
       "4     log   p0 -4.070879e+01 -8.063180e+01 -7.857815e-01   \n",
       "5     log   p1  1.252720e-08 -5.050257e-09  3.010466e-08   \n",
       "6     log   p2  1.507243e+01  6.671581e+00  2.347327e+01   \n",
       "7     log   p3 -1.236510e+01 -1.864359e+01 -6.086617e+00   \n",
       "8   plain   p0 -1.239927e+02 -2.339494e+02 -9.562457e+00   \n",
       "9   plain   p1 -4.315575e-05 -9.595711e-05  7.713982e-06   \n",
       "10  plain   p2  5.582677e+00  2.794437e+00  8.270209e+00   \n",
       "11  plain   p3 -1.279821e+01 -1.863649e+01 -6.719472e+00   \n",
       "12  plain   p0 -1.239927e+02 -6.620773e+02 -2.093170e+01   \n",
       "13  plain   p1 -4.315575e-05 -2.420179e+00  3.871659e-04   \n",
       "14  plain   p2  5.582677e+00 -1.102782e+01  1.781766e+01   \n",
       "15  plain   p3 -1.279821e+01 -3.283880e+01  5.789172e+00   \n",
       "16  plain   p0 -1.239927e+02 -6.477410e+02 -2.141250e+01   \n",
       "17  plain   p1 -4.315575e-05 -2.530518e+00  2.107221e-01   \n",
       "18  plain   p2  5.582677e+00 -1.103487e+01  1.753722e+01   \n",
       "19  plain   p3 -1.279821e+01 -3.243664e+01  5.801690e+00   \n",
       "20  plain   p0 -1.239927e+02 -5.199186e+02 -1.426793e+01   \n",
       "21  plain   p1 -4.315575e-05 -2.558559e+00  1.758763e-02   \n",
       "22  plain   p2  5.582677e+00 -1.729444e+01  1.472734e+01   \n",
       "23  plain   p3 -1.279821e+01 -2.808349e+01  2.641113e+01   \n",
       "24    log   p0 -4.070879e+01 -8.051482e+01  5.768097e-01   \n",
       "25    log   p1  1.252720e-08 -4.475513e-09  2.942177e-08   \n",
       "26    log   p2  1.507243e+01  6.428671e+00  2.346898e+01   \n",
       "27    log   p3 -1.236510e+01 -1.860056e+01 -5.869817e+00   \n",
       "28    log   p0 -4.070879e+01 -3.623419e+02 -1.445765e+01   \n",
       "29    log   p1  1.252720e-08 -1.294969e-05  1.370159e-05   \n",
       "30    log   p2  1.507243e+01 -3.574117e+01  7.860083e+01   \n",
       "31    log   p3 -1.236510e+01 -4.406300e+01  2.279963e+00   \n",
       "32    log   p0 -4.070879e+01 -3.693511e+02 -1.420581e+01   \n",
       "33    log   p1  1.252720e-08 -1.140706e-05  1.209935e-05   \n",
       "34    log   p2  1.507243e+01 -3.575099e+01  7.927186e+01   \n",
       "35    log   p3 -1.236510e+01 -4.449996e+01  2.280283e+00   \n",
       "36    log   p0 -4.070879e+01 -3.109718e+02 -8.945695e+00   \n",
       "37    log   p1  1.252720e-08 -2.203192e-04  2.410079e-04   \n",
       "38    log   p2  1.507243e+01 -7.406005e+01  4.862000e+01   \n",
       "39    log   p3 -1.236510e+01 -3.497653e+01  1.768902e+01   \n",
       "\n",
       "                       method  \n",
       "0                  Wald/delta  \n",
       "1                  Wald/delta  \n",
       "2                  Wald/delta  \n",
       "3                  Wald/delta  \n",
       "4                  Wald/delta  \n",
       "5                  Wald/delta  \n",
       "6                  Wald/delta  \n",
       "7                  Wald/delta  \n",
       "8                      normal  \n",
       "9                      normal  \n",
       "10                     normal  \n",
       "11                     normal  \n",
       "12             nonparam_boots  \n",
       "13             nonparam_boots  \n",
       "14             nonparam_boots  \n",
       "15             nonparam_boots  \n",
       "16  nonparam_stratified_boots  \n",
       "17  nonparam_stratified_boots  \n",
       "18  nonparam_stratified_boots  \n",
       "19  nonparam_stratified_boots  \n",
       "20           parametric_boots  \n",
       "21           parametric_boots  \n",
       "22           parametric_boots  \n",
       "23           parametric_boots  \n",
       "24                     normal  \n",
       "25                     normal  \n",
       "26                     normal  \n",
       "27                     normal  \n",
       "28             nonparam_boots  \n",
       "29             nonparam_boots  \n",
       "30             nonparam_boots  \n",
       "31             nonparam_boots  \n",
       "32  nonparam_stratified_boots  \n",
       "33  nonparam_stratified_boots  \n",
       "34  nonparam_stratified_boots  \n",
       "35  nonparam_stratified_boots  \n",
       "36           parametric_boots  \n",
       "37           parametric_boots  \n",
       "38           parametric_boots  \n",
       "39           parametric_boots  "
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# nr of parameters\n",
    "d = thetas.shape[-1]\n",
    "\n",
    "# CI of params:\n",
    "#   - using Wald approximation \n",
    "#   - assuming asymptotic MLE distr of parameters \n",
    "#   - cov = hessian(nllf)^{-1}\n",
    "#   - this could be considered as delta method\n",
    "\n",
    "print(f\"theta CI:scale:{scales},method:Wald/delta\")\n",
    "lst_theta_CI = [df_theta_CI_wald.assign(method= \"Wald/delta\")]\n",
    "\n",
    "for scale, theta in zip(scales, thetas):\n",
    "\n",
    "    # parameter quantiles from sampled or bootstrapped parameter sets\n",
    "    for method in [\"normal\", \"nonparam_boots\", \"nonparam_stratified_boots\", \"parametric_boots\"]:\n",
    "    \n",
    "        print(f\"model CI:scale:{scale},boots-method:{method}\")\n",
    "\n",
    "        # get sampled or bootstrapped parameters\n",
    "        btheta = boots_theta_results[(scale, method)]\n",
    "    \n",
    "        # quantiles\n",
    "        quant = np.quantile(btheta, probs, axis = 0)\n",
    "\n",
    "        df_tmp = pd.DataFrame({\n",
    "            \"method\": f\"{method}\",\n",
    "            \"scale\" : scale,\n",
    "            \"coef\" : [f\"p{i}\" for i in range(d)],\n",
    "            \"value\" : theta, \n",
    "            \"LCL\" : quant[0], \n",
    "            \"UCL\": quant[1]\n",
    "        })\n",
    "        \n",
    "        lst_theta_CI.append(df_tmp)\n",
    "\n",
    "df_theta_CI = pd.concat(lst_theta_CI, ignore_index=True)\n",
    "df_theta_CI"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "86e2290df4d1",
   "metadata": {},
   "source": [
    "### 4.6 Beta-parameter confidence-interval comparison\n",
    "\n",
    "Here $\\beta=(\\beta_0,\\beta_1,\\beta_2,\\beta_3)$ denotes the coefficients of the polynomial logit $\\eta(x)=\\sum_j\\beta_jx^j$. For the monotonic cubic model, each complete sampled or bootstrapped $\\theta$ vector is first mapped with `get_beta`; coefficient quantiles are calculated only after this transformation. Transforming the endpoints of marginal $\\theta$ intervals would not be valid because the map $\\theta\\mapsto\\beta$ is nonlinear.\n",
    "\n",
    "The table includes the unconstrained cubic Wald/delta result (where $\\beta=\\theta$) and the monotonic cubic Wald/delta, normal-sampling, and bootstrap results. `normal` propagates draws from the joint normal approximation through the nonlinear map, while `Wald/delta` uses the first-order covariance $J_{\\beta\\theta}\\,\\mathrm{Cov}(\\theta)\\,J_{\\beta\\theta}^{T}$. The displayed intervals are marginal, not joint.\n",
    "\n",
    "The $\\beta$ coefficients are invariant to the sign symmetries that make the monotonic $\\theta$ representation non-unique, so they are the more interpretable parameters. Their numerical values still depend on the predictor scale and origin: the `plain` and `log` columns provide a side-by-side comparison of separately defined coefficient systems, not coefficients that should be numerically equal.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "99a3633506aa",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "index",
         "rawType": "int64",
         "type": "integer"
        },
        {
         "name": "model",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "scale",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "method",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "coef",
         "rawType": "object",
         "type": "string"
        },
        {
         "name": "value",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "LCL",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "UCL",
         "rawType": "float64",
         "type": "float"
        }
       ],
       "ref": "f1acbc6b-2096-4e94-a9bd-5fb9db482129",
       "rows": [
        [
         "0",
         "cubic unconstrained",
         "plain",
         "Wald/delta",
         "beta_0",
         "-197.43456924470524",
         "-402.4329780401599",
         "7.56383955074935"
        ],
        [
         "1",
         "cubic unconstrained",
         "plain",
         "Wald/delta",
         "beta_1",
         "248.4905000484067",
         "-14.566758126806093",
         "511.54775822361944"
        ],
        [
         "2",
         "cubic unconstrained",
         "plain",
         "Wald/delta",
         "beta_2",
         "-98.59976324737492",
         "-205.4692297024859",
         "8.269703207736043"
        ],
        [
         "3",
         "cubic unconstrained",
         "plain",
         "Wald/delta",
         "beta_3",
         "12.18035307606018",
         "-1.418592629226529",
         "25.77929878134689"
        ],
        [
         "4",
         "cubic unconstrained",
         "log",
         "Wald/delta",
         "beta_0",
         "-75.33868079212766",
         "-162.96251965553296",
         "12.285158071277635"
        ],
        [
         "5",
         "cubic unconstrained",
         "log",
         "Wald/delta",
         "beta_1",
         "287.5173164621367",
         "-51.62426932699947",
         "626.6589022512728"
        ],
        [
         "6",
         "cubic unconstrained",
         "log",
         "Wald/delta",
         "beta_2",
         "-339.6550928412925",
         "-748.8691888972139",
         "69.55900321462877"
        ],
        [
         "7",
         "cubic unconstrained",
         "log",
         "Wald/delta",
         "beta_3",
         "124.51434906316405",
         "-29.18318823491893",
         "278.211886361247"
        ],
        [
         "8",
         "cubic monotonic",
         "plain",
         "Wald/delta",
         "beta_0",
         "-123.99266532844526",
         "-234.62067941579588",
         "-13.364651241094663"
        ],
        [
         "9",
         "cubic monotonic",
         "plain",
         "Wald/delta",
         "beta_1",
         "163.79429026354094",
         "14.050524053789502",
         "313.5380564732924"
        ],
        [
         "10",
         "cubic monotonic",
         "plain",
         "Wald/delta",
         "beta_2",
         "-71.44830230346912",
         "-138.5378636511295",
         "-4.358740955808756"
        ],
        [
         "11",
         "cubic monotonic",
         "plain",
         "Wald/delta",
         "beta_3",
         "10.38876242041641",
         "0.34507665364804296",
         "20.432448187184775"
        ],
        [
         "12",
         "cubic monotonic",
         "plain",
         "normal",
         "beta_0",
         "-123.99266532844526",
         "-233.94943123214563",
         "-9.562456807989498"
        ],
        [
         "13",
         "cubic monotonic",
         "plain",
         "normal",
         "beta_1",
         "163.79429026354094",
         "45.151308613356704",
         "347.31875171406745"
        ],
        [
         "14",
         "cubic monotonic",
         "plain",
         "normal",
         "beta_2",
         "-71.44830230346912",
         "-153.87750482778767",
         "-18.65154213210847"
        ],
        [
         "15",
         "cubic monotonic",
         "plain",
         "normal",
         "beta_3",
         "10.38876242041641",
         "2.602958679422927",
         "22.79878634431914"
        ],
        [
         "16",
         "cubic monotonic",
         "plain",
         "nonparam_boots",
         "beta_0",
         "-123.99266532844526",
         "-662.077284728864",
         "-20.9316993950732"
        ],
        [
         "17",
         "cubic monotonic",
         "plain",
         "nonparam_boots",
         "beta_1",
         "163.79429026354094",
         "7.451974330719892",
         "1082.6209802364356"
        ],
        [
         "18",
         "cubic monotonic",
         "plain",
         "nonparam_boots",
         "beta_2",
         "-71.44830230346912",
         "-586.9564245512602",
         "-3.339262998205471"
        ],
        [
         "19",
         "cubic monotonic",
         "plain",
         "nonparam_boots",
         "beta_3",
         "10.38876242041641",
         "0.7696558440480198",
         "108.79161500766871"
        ],
        [
         "20",
         "cubic monotonic",
         "plain",
         "nonparam_stratified_boots",
         "beta_0",
         "-123.99266532844526",
         "-647.740989994033",
         "-21.412504234536605"
        ],
        [
         "21",
         "cubic monotonic",
         "plain",
         "nonparam_stratified_boots",
         "beta_1",
         "163.79429026354094",
         "7.756705269203346",
         "1057.0439934366461"
        ],
        [
         "22",
         "cubic monotonic",
         "plain",
         "nonparam_stratified_boots",
         "beta_2",
         "-71.44830230346912",
         "-576.9531658776733",
         "-13.715893635519103"
        ],
        [
         "23",
         "cubic monotonic",
         "plain",
         "nonparam_stratified_boots",
         "beta_3",
         "10.38876242041641",
         "6.602426024917408",
         "106.65611161633856"
        ],
        [
         "24",
         "cubic monotonic",
         "plain",
         "parametric_boots",
         "beta_0",
         "-123.99266532844526",
         "-519.9185959807692",
         "-14.267927545473208"
        ],
        [
         "25",
         "cubic monotonic",
         "plain",
         "parametric_boots",
         "beta_1",
         "163.79429026354094",
         "0.2691424410169285",
         "858.6380608133204"
        ],
        [
         "26",
         "cubic monotonic",
         "plain",
         "parametric_boots",
         "beta_2",
         "-71.44830230346912",
         "-504.895599785272",
         "-1.3044258940089664"
        ],
        [
         "27",
         "cubic monotonic",
         "plain",
         "parametric_boots",
         "beta_3",
         "10.38876242041641",
         "2.109768489014601",
         "101.82745503517135"
        ],
        [
         "28",
         "cubic monotonic",
         "log",
         "Wald/delta",
         "beta_0",
         "-40.70878922895756",
         "-80.63179697227875",
         "-0.7857814856363916"
        ],
        [
         "29",
         "cubic monotonic",
         "log",
         "Wald/delta",
         "beta_1",
         "152.8957509914481",
         "-2.372459765504317",
         "308.1639617484005"
        ],
        [
         "30",
         "cubic monotonic",
         "log",
         "Wald/delta",
         "beta_2",
         "-186.3720915226409",
         "-384.1560897326805",
         "11.411906687398698"
        ],
        [
         "31",
         "cubic monotonic",
         "log",
         "Wald/delta",
         "beta_3",
         "75.72601216469042",
         "-8.688065429465865",
         "160.14008975884667"
        ],
        [
         "32",
         "cubic monotonic",
         "log",
         "normal",
         "beta_0",
         "-40.70878922895756",
         "-80.51482391483258",
         "0.5768096898392844"
        ],
        [
         "33",
         "cubic monotonic",
         "log",
         "normal",
         "beta_1",
         "152.8957509914481",
         "34.45474859990287",
         "345.9810094360133"
        ],
        [
         "34",
         "cubic monotonic",
         "log",
         "normal",
         "beta_2",
         "-186.3720915226409",
         "-434.56515423468676",
         "-38.043108560505516"
        ],
        [
         "35",
         "cubic monotonic",
         "log",
         "normal",
         "beta_3",
         "75.72601216469042",
         "13.775937973930814",
         "183.59771002934062"
        ],
        [
         "36",
         "cubic monotonic",
         "log",
         "nonparam_boots",
         "beta_0",
         "-40.70878922895756",
         "-362.3418986772154",
         "-14.457646430958098"
        ],
        [
         "37",
         "cubic monotonic",
         "log",
         "nonparam_boots",
         "beta_1",
         "152.8957509914481",
         "0.008195165685316728",
         "1941.5479181241703"
        ],
        [
         "38",
         "cubic monotonic",
         "log",
         "nonparam_boots",
         "beta_2",
         "-186.3720915226409",
         "-3466.7015262176637",
         "2.548175943669855"
        ],
        [
         "39",
         "cubic monotonic",
         "log",
         "nonparam_boots",
         "beta_3",
         "75.72601216469042",
         "2.0700275975124494",
         "2059.363688406808"
        ],
        [
         "40",
         "cubic monotonic",
         "log",
         "nonparam_stratified_boots",
         "beta_0",
         "-40.70878922895756",
         "-369.35112799570214",
         "-14.205808684448709"
        ],
        [
         "41",
         "cubic monotonic",
         "log",
         "nonparam_stratified_boots",
         "beta_1",
         "152.8957509914481",
         "0.01446377082647642",
         "1980.2467211492028"
        ],
        [
         "42",
         "cubic monotonic",
         "log",
         "nonparam_stratified_boots",
         "beta_2",
         "-186.3720915226409",
         "-3528.049143561927",
         "2.305802812494407"
        ],
        [
         "43",
         "cubic monotonic",
         "log",
         "nonparam_stratified_boots",
         "beta_3",
         "75.72601216469042",
         "1.7517826147937314",
         "2094.6757464165717"
        ],
        [
         "44",
         "cubic monotonic",
         "log",
         "parametric_boots",
         "beta_0",
         "-40.70878922895756",
         "-310.97179686469656",
         "-8.945695488818114"
        ],
        [
         "45",
         "cubic monotonic",
         "log",
         "parametric_boots",
         "beta_1",
         "152.8957509914481",
         "0.1349783013815043",
         "1223.3577524191335"
        ],
        [
         "46",
         "cubic monotonic",
         "log",
         "parametric_boots",
         "beta_2",
         "-186.3720915226409",
         "-1600.4542962428347",
         "6.0867605543462355"
        ],
        [
         "47",
         "cubic monotonic",
         "log",
         "parametric_boots",
         "beta_3",
         "75.72601216469042",
         "0.9841053188719839",
         "1828.2970471267927"
        ]
       ],
       "shape": {
        "columns": 7,
        "rows": 48
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>model</th>\n",
       "      <th>scale</th>\n",
       "      <th>method</th>\n",
       "      <th>coef</th>\n",
       "      <th>value</th>\n",
       "      <th>LCL</th>\n",
       "      <th>UCL</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>cubic unconstrained</td>\n",
       "      <td>plain</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-197.434569</td>\n",
       "      <td>-402.432978</td>\n",
       "      <td>7.563840</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>cubic unconstrained</td>\n",
       "      <td>plain</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>248.490500</td>\n",
       "      <td>-14.566758</td>\n",
       "      <td>511.547758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>cubic unconstrained</td>\n",
       "      <td>plain</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-98.599763</td>\n",
       "      <td>-205.469230</td>\n",
       "      <td>8.269703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>cubic unconstrained</td>\n",
       "      <td>plain</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>12.180353</td>\n",
       "      <td>-1.418593</td>\n",
       "      <td>25.779299</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>cubic unconstrained</td>\n",
       "      <td>log</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-75.338681</td>\n",
       "      <td>-162.962520</td>\n",
       "      <td>12.285158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>cubic unconstrained</td>\n",
       "      <td>log</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>287.517316</td>\n",
       "      <td>-51.624269</td>\n",
       "      <td>626.658902</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>cubic unconstrained</td>\n",
       "      <td>log</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-339.655093</td>\n",
       "      <td>-748.869189</td>\n",
       "      <td>69.559003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>cubic unconstrained</td>\n",
       "      <td>log</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>124.514349</td>\n",
       "      <td>-29.183188</td>\n",
       "      <td>278.211886</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-234.620679</td>\n",
       "      <td>-13.364651</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>163.794290</td>\n",
       "      <td>14.050524</td>\n",
       "      <td>313.538056</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-71.448302</td>\n",
       "      <td>-138.537864</td>\n",
       "      <td>-4.358741</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>10.388762</td>\n",
       "      <td>0.345077</td>\n",
       "      <td>20.432448</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>normal</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-233.949431</td>\n",
       "      <td>-9.562457</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>normal</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>163.794290</td>\n",
       "      <td>45.151309</td>\n",
       "      <td>347.318752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>normal</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-71.448302</td>\n",
       "      <td>-153.877505</td>\n",
       "      <td>-18.651542</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>normal</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>10.388762</td>\n",
       "      <td>2.602959</td>\n",
       "      <td>22.798786</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>nonparam_boots</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-662.077285</td>\n",
       "      <td>-20.931699</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>nonparam_boots</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>163.794290</td>\n",
       "      <td>7.451974</td>\n",
       "      <td>1082.620980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>nonparam_boots</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-71.448302</td>\n",
       "      <td>-586.956425</td>\n",
       "      <td>-3.339263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>nonparam_boots</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>10.388762</td>\n",
       "      <td>0.769656</td>\n",
       "      <td>108.791615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-647.740990</td>\n",
       "      <td>-21.412504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>163.794290</td>\n",
       "      <td>7.756705</td>\n",
       "      <td>1057.043993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-71.448302</td>\n",
       "      <td>-576.953166</td>\n",
       "      <td>-13.715894</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>10.388762</td>\n",
       "      <td>6.602426</td>\n",
       "      <td>106.656112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>parametric_boots</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-519.918596</td>\n",
       "      <td>-14.267928</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>parametric_boots</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>163.794290</td>\n",
       "      <td>0.269142</td>\n",
       "      <td>858.638061</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>parametric_boots</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-71.448302</td>\n",
       "      <td>-504.895600</td>\n",
       "      <td>-1.304426</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>plain</td>\n",
       "      <td>parametric_boots</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>10.388762</td>\n",
       "      <td>2.109768</td>\n",
       "      <td>101.827455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-40.708789</td>\n",
       "      <td>-80.631797</td>\n",
       "      <td>-0.785781</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>152.895751</td>\n",
       "      <td>-2.372460</td>\n",
       "      <td>308.163962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-186.372092</td>\n",
       "      <td>-384.156090</td>\n",
       "      <td>11.411907</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>Wald/delta</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>75.726012</td>\n",
       "      <td>-8.688065</td>\n",
       "      <td>160.140090</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>normal</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-40.708789</td>\n",
       "      <td>-80.514824</td>\n",
       "      <td>0.576810</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>normal</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>152.895751</td>\n",
       "      <td>34.454749</td>\n",
       "      <td>345.981009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>normal</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-186.372092</td>\n",
       "      <td>-434.565154</td>\n",
       "      <td>-38.043109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>normal</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>75.726012</td>\n",
       "      <td>13.775938</td>\n",
       "      <td>183.597710</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>nonparam_boots</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-40.708789</td>\n",
       "      <td>-362.341899</td>\n",
       "      <td>-14.457646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>nonparam_boots</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>152.895751</td>\n",
       "      <td>0.008195</td>\n",
       "      <td>1941.547918</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>nonparam_boots</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-186.372092</td>\n",
       "      <td>-3466.701526</td>\n",
       "      <td>2.548176</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>nonparam_boots</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>75.726012</td>\n",
       "      <td>2.070028</td>\n",
       "      <td>2059.363688</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-40.708789</td>\n",
       "      <td>-369.351128</td>\n",
       "      <td>-14.205809</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>152.895751</td>\n",
       "      <td>0.014464</td>\n",
       "      <td>1980.246721</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-186.372092</td>\n",
       "      <td>-3528.049144</td>\n",
       "      <td>2.305803</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>nonparam_stratified_boots</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>75.726012</td>\n",
       "      <td>1.751783</td>\n",
       "      <td>2094.675746</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>parametric_boots</td>\n",
       "      <td>beta_0</td>\n",
       "      <td>-40.708789</td>\n",
       "      <td>-310.971797</td>\n",
       "      <td>-8.945695</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>parametric_boots</td>\n",
       "      <td>beta_1</td>\n",
       "      <td>152.895751</td>\n",
       "      <td>0.134978</td>\n",
       "      <td>1223.357752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>parametric_boots</td>\n",
       "      <td>beta_2</td>\n",
       "      <td>-186.372092</td>\n",
       "      <td>-1600.454296</td>\n",
       "      <td>6.086761</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>cubic monotonic</td>\n",
       "      <td>log</td>\n",
       "      <td>parametric_boots</td>\n",
       "      <td>beta_3</td>\n",
       "      <td>75.726012</td>\n",
       "      <td>0.984105</td>\n",
       "      <td>1828.297047</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  model  scale                     method    coef       value  \\\n",
       "0   cubic unconstrained  plain                 Wald/delta  beta_0 -197.434569   \n",
       "1   cubic unconstrained  plain                 Wald/delta  beta_1  248.490500   \n",
       "2   cubic unconstrained  plain                 Wald/delta  beta_2  -98.599763   \n",
       "3   cubic unconstrained  plain                 Wald/delta  beta_3   12.180353   \n",
       "4   cubic unconstrained    log                 Wald/delta  beta_0  -75.338681   \n",
       "5   cubic unconstrained    log                 Wald/delta  beta_1  287.517316   \n",
       "6   cubic unconstrained    log                 Wald/delta  beta_2 -339.655093   \n",
       "7   cubic unconstrained    log                 Wald/delta  beta_3  124.514349   \n",
       "8       cubic monotonic  plain                 Wald/delta  beta_0 -123.992665   \n",
       "9       cubic monotonic  plain                 Wald/delta  beta_1  163.794290   \n",
       "10      cubic monotonic  plain                 Wald/delta  beta_2  -71.448302   \n",
       "11      cubic monotonic  plain                 Wald/delta  beta_3   10.388762   \n",
       "12      cubic monotonic  plain                     normal  beta_0 -123.992665   \n",
       "13      cubic monotonic  plain                     normal  beta_1  163.794290   \n",
       "14      cubic monotonic  plain                     normal  beta_2  -71.448302   \n",
       "15      cubic monotonic  plain                     normal  beta_3   10.388762   \n",
       "16      cubic monotonic  plain             nonparam_boots  beta_0 -123.992665   \n",
       "17      cubic monotonic  plain             nonparam_boots  beta_1  163.794290   \n",
       "18      cubic monotonic  plain             nonparam_boots  beta_2  -71.448302   \n",
       "19      cubic monotonic  plain             nonparam_boots  beta_3   10.388762   \n",
       "20      cubic monotonic  plain  nonparam_stratified_boots  beta_0 -123.992665   \n",
       "21      cubic monotonic  plain  nonparam_stratified_boots  beta_1  163.794290   \n",
       "22      cubic monotonic  plain  nonparam_stratified_boots  beta_2  -71.448302   \n",
       "23      cubic monotonic  plain  nonparam_stratified_boots  beta_3   10.388762   \n",
       "24      cubic monotonic  plain           parametric_boots  beta_0 -123.992665   \n",
       "25      cubic monotonic  plain           parametric_boots  beta_1  163.794290   \n",
       "26      cubic monotonic  plain           parametric_boots  beta_2  -71.448302   \n",
       "27      cubic monotonic  plain           parametric_boots  beta_3   10.388762   \n",
       "28      cubic monotonic    log                 Wald/delta  beta_0  -40.708789   \n",
       "29      cubic monotonic    log                 Wald/delta  beta_1  152.895751   \n",
       "30      cubic monotonic    log                 Wald/delta  beta_2 -186.372092   \n",
       "31      cubic monotonic    log                 Wald/delta  beta_3   75.726012   \n",
       "32      cubic monotonic    log                     normal  beta_0  -40.708789   \n",
       "33      cubic monotonic    log                     normal  beta_1  152.895751   \n",
       "34      cubic monotonic    log                     normal  beta_2 -186.372092   \n",
       "35      cubic monotonic    log                     normal  beta_3   75.726012   \n",
       "36      cubic monotonic    log             nonparam_boots  beta_0  -40.708789   \n",
       "37      cubic monotonic    log             nonparam_boots  beta_1  152.895751   \n",
       "38      cubic monotonic    log             nonparam_boots  beta_2 -186.372092   \n",
       "39      cubic monotonic    log             nonparam_boots  beta_3   75.726012   \n",
       "40      cubic monotonic    log  nonparam_stratified_boots  beta_0  -40.708789   \n",
       "41      cubic monotonic    log  nonparam_stratified_boots  beta_1  152.895751   \n",
       "42      cubic monotonic    log  nonparam_stratified_boots  beta_2 -186.372092   \n",
       "43      cubic monotonic    log  nonparam_stratified_boots  beta_3   75.726012   \n",
       "44      cubic monotonic    log           parametric_boots  beta_0  -40.708789   \n",
       "45      cubic monotonic    log           parametric_boots  beta_1  152.895751   \n",
       "46      cubic monotonic    log           parametric_boots  beta_2 -186.372092   \n",
       "47      cubic monotonic    log           parametric_boots  beta_3   75.726012   \n",
       "\n",
       "            LCL          UCL  \n",
       "0   -402.432978     7.563840  \n",
       "1    -14.566758   511.547758  \n",
       "2   -205.469230     8.269703  \n",
       "3     -1.418593    25.779299  \n",
       "4   -162.962520    12.285158  \n",
       "5    -51.624269   626.658902  \n",
       "6   -748.869189    69.559003  \n",
       "7    -29.183188   278.211886  \n",
       "8   -234.620679   -13.364651  \n",
       "9     14.050524   313.538056  \n",
       "10  -138.537864    -4.358741  \n",
       "11     0.345077    20.432448  \n",
       "12  -233.949431    -9.562457  \n",
       "13    45.151309   347.318752  \n",
       "14  -153.877505   -18.651542  \n",
       "15     2.602959    22.798786  \n",
       "16  -662.077285   -20.931699  \n",
       "17     7.451974  1082.620980  \n",
       "18  -586.956425    -3.339263  \n",
       "19     0.769656   108.791615  \n",
       "20  -647.740990   -21.412504  \n",
       "21     7.756705  1057.043993  \n",
       "22  -576.953166   -13.715894  \n",
       "23     6.602426   106.656112  \n",
       "24  -519.918596   -14.267928  \n",
       "25     0.269142   858.638061  \n",
       "26  -504.895600    -1.304426  \n",
       "27     2.109768   101.827455  \n",
       "28   -80.631797    -0.785781  \n",
       "29    -2.372460   308.163962  \n",
       "30  -384.156090    11.411907  \n",
       "31    -8.688065   160.140090  \n",
       "32   -80.514824     0.576810  \n",
       "33    34.454749   345.981009  \n",
       "34  -434.565154   -38.043109  \n",
       "35    13.775938   183.597710  \n",
       "36  -362.341899   -14.457646  \n",
       "37     0.008195  1941.547918  \n",
       "38 -3466.701526     2.548176  \n",
       "39     2.070028  2059.363688  \n",
       "40  -369.351128   -14.205809  \n",
       "41     0.014464  1980.246721  \n",
       "42 -3528.049144     2.305803  \n",
       "43     1.751783  2094.675746  \n",
       "44  -310.971797    -8.945695  \n",
       "45     0.134978  1223.357752  \n",
       "46 -1600.454296     6.086761  \n",
       "47     0.984105  1828.297047  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "beta_ci_rows = []\n",
    "\n",
    "def add_beta_ci_rows(model_name, scale, method, beta, limits):\n",
    "    for index, (estimate, lower, upper) in enumerate(\n",
    "        zip(beta, limits[0], limits[1])\n",
    "    ):\n",
    "        beta_ci_rows.append({\n",
    "            \"model\": model_name,\n",
    "            \"scale\": scale,\n",
    "            \"method\": method,\n",
    "            \"coef\": f\"beta_{index}\",\n",
    "            \"value\": estimate,\n",
    "            \"LCL\": lower,\n",
    "            \"UCL\": upper,\n",
    "        })\n",
    "\n",
    "# Unconstrained cubic: beta = theta, so only the already-computed\n",
    "# Wald/delta result is included.\n",
    "unconstrained_model = cubic_unconstrained_results[\"model\"]\n",
    "for scale, theta, cov_theta in zip(\n",
    "    scales,\n",
    "    cubic_unconstrained_results[\"parameters\"],\n",
    "    cubic_unconstrained_results[\"covariances\"],\n",
    "):\n",
    "    beta = unconstrained_model.get_beta(theta)\n",
    "    jac_beta = unconstrained_model.get_jac_beta(theta)\n",
    "    cov_beta = jac_beta @ cov_theta @ jac_beta.T\n",
    "    limits = unconstrained_model.get_theta_quantiles_normal(\n",
    "        probs, beta, cov_beta\n",
    "    )\n",
    "    add_beta_ci_rows(\"cubic unconstrained\", scale, \"Wald/delta\", beta, limits)\n",
    "\n",
    "# Monotonic cubic: compare the first-order delta result with the\n",
    "# nonlinear transformation of all normal-sampled/bootstrap draws.\n",
    "monotonic_model = logit\n",
    "parameter_methods = [\n",
    "    \"normal\",\n",
    "    \"nonparam_boots\",\n",
    "    \"nonparam_stratified_boots\",\n",
    "    \"parametric_boots\",\n",
    "]\n",
    "\n",
    "for scale, theta, cov_theta in zip(scales, thetas, cov_thetas):\n",
    "    beta = monotonic_model.get_beta(theta)\n",
    "    jac_beta = monotonic_model.get_jac_beta(theta)\n",
    "    cov_beta = jac_beta @ cov_theta @ jac_beta.T\n",
    "    limits = monotonic_model.get_theta_quantiles_normal(probs, beta, cov_beta)\n",
    "    add_beta_ci_rows(\"cubic monotonic\", scale, \"Wald/delta\", beta, limits)\n",
    "\n",
    "    for method in parameter_methods:\n",
    "        theta_draws = boots_theta_results[(scale, method)]\n",
    "        beta_draws = np.asarray([\n",
    "            monotonic_model.get_beta(theta_draw)\n",
    "            for theta_draw in theta_draws\n",
    "        ])\n",
    "        limits = np.quantile(beta_draws, probs, axis=0)\n",
    "        add_beta_ci_rows(\"cubic monotonic\", scale, method, beta, limits)\n",
    "\n",
    "df_beta_CI = pd.DataFrame(beta_ci_rows)\n",
    "cubic_unconstrained_results[\"beta_parameter_ci_comparison\"] = (\n",
    "    df_beta_CI.loc[df_beta_CI[\"model\"] == \"cubic unconstrained\"]\n",
    "    .reset_index(drop=True)\n",
    ")\n",
    "df_beta_CI\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "14610b3689ac",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.microsoft.datawrangler.viewer.v0+json": {
       "columns": [
        {
         "name": "('model', 'coef', 'method')",
         "rawType": "object",
         "type": "unknown"
        },
        {
         "name": "('value', 'log')",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "('value', 'plain')",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "('LCL', 'log')",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "('LCL', 'plain')",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "('UCL', 'log')",
         "rawType": "float64",
         "type": "float"
        },
        {
         "name": "('UCL', 'plain')",
         "rawType": "float64",
         "type": "float"
        }
       ],
       "ref": "a15d452a-ca78-4c8d-9d58-cccec1ba556b",
       "rows": [
        [
         "('cubic monotonic', 'beta_0', 'Wald/delta')",
         "-40.70878922895756",
         "-123.99266532844526",
         "-80.63179697227875",
         "-234.62067941579588",
         "-0.7857814856363916",
         "-13.364651241094663"
        ],
        [
         "('cubic monotonic', 'beta_0', 'nonparam_boots')",
         "-40.70878922895756",
         "-123.99266532844526",
         "-362.3418986772154",
         "-662.077284728864",
         "-14.457646430958098",
         "-20.9316993950732"
        ],
        [
         "('cubic monotonic', 'beta_0', 'nonparam_stratified_boots')",
         "-40.70878922895756",
         "-123.99266532844526",
         "-369.35112799570214",
         "-647.740989994033",
         "-14.205808684448709",
         "-21.412504234536605"
        ],
        [
         "('cubic monotonic', 'beta_0', 'normal')",
         "-40.70878922895756",
         "-123.99266532844526",
         "-80.51482391483258",
         "-233.94943123214563",
         "0.5768096898392844",
         "-9.562456807989498"
        ],
        [
         "('cubic monotonic', 'beta_0', 'parametric_boots')",
         "-40.70878922895756",
         "-123.99266532844526",
         "-310.97179686469656",
         "-519.9185959807692",
         "-8.945695488818114",
         "-14.267927545473208"
        ],
        [
         "('cubic monotonic', 'beta_1', 'Wald/delta')",
         "152.8957509914481",
         "163.79429026354094",
         "-2.372459765504317",
         "14.050524053789502",
         "308.1639617484005",
         "313.5380564732924"
        ],
        [
         "('cubic monotonic', 'beta_1', 'nonparam_boots')",
         "152.8957509914481",
         "163.79429026354094",
         "0.008195165685316728",
         "7.451974330719892",
         "1941.5479181241703",
         "1082.6209802364356"
        ],
        [
         "('cubic monotonic', 'beta_1', 'nonparam_stratified_boots')",
         "152.8957509914481",
         "163.79429026354094",
         "0.01446377082647642",
         "7.756705269203346",
         "1980.2467211492028",
         "1057.0439934366461"
        ],
        [
         "('cubic monotonic', 'beta_1', 'normal')",
         "152.8957509914481",
         "163.79429026354094",
         "34.45474859990287",
         "45.151308613356704",
         "345.9810094360133",
         "347.31875171406745"
        ],
        [
         "('cubic monotonic', 'beta_1', 'parametric_boots')",
         "152.8957509914481",
         "163.79429026354094",
         "0.1349783013815043",
         "0.2691424410169285",
         "1223.3577524191335",
         "858.6380608133204"
        ],
        [
         "('cubic monotonic', 'beta_2', 'Wald/delta')",
         "-186.3720915226409",
         "-71.44830230346912",
         "-384.1560897326805",
         "-138.5378636511295",
         "11.411906687398698",
         "-4.358740955808756"
        ],
        [
         "('cubic monotonic', 'beta_2', 'nonparam_boots')",
         "-186.3720915226409",
         "-71.44830230346912",
         "-3466.7015262176637",
         "-586.9564245512602",
         "2.548175943669855",
         "-3.339262998205471"
        ],
        [
         "('cubic monotonic', 'beta_2', 'nonparam_stratified_boots')",
         "-186.3720915226409",
         "-71.44830230346912",
         "-3528.049143561927",
         "-576.9531658776733",
         "2.305802812494407",
         "-13.715893635519103"
        ],
        [
         "('cubic monotonic', 'beta_2', 'normal')",
         "-186.3720915226409",
         "-71.44830230346912",
         "-434.56515423468676",
         "-153.87750482778767",
         "-38.043108560505516",
         "-18.65154213210847"
        ],
        [
         "('cubic monotonic', 'beta_2', 'parametric_boots')",
         "-186.3720915226409",
         "-71.44830230346912",
         "-1600.4542962428347",
         "-504.895599785272",
         "6.0867605543462355",
         "-1.3044258940089664"
        ],
        [
         "('cubic monotonic', 'beta_3', 'Wald/delta')",
         "75.72601216469042",
         "10.38876242041641",
         "-8.688065429465865",
         "0.34507665364804296",
         "160.14008975884667",
         "20.432448187184775"
        ],
        [
         "('cubic monotonic', 'beta_3', 'nonparam_boots')",
         "75.72601216469042",
         "10.38876242041641",
         "2.0700275975124494",
         "0.7696558440480198",
         "2059.363688406808",
         "108.79161500766871"
        ],
        [
         "('cubic monotonic', 'beta_3', 'nonparam_stratified_boots')",
         "75.72601216469042",
         "10.38876242041641",
         "1.7517826147937314",
         "6.602426024917408",
         "2094.6757464165717",
         "106.65611161633856"
        ],
        [
         "('cubic monotonic', 'beta_3', 'normal')",
         "75.72601216469042",
         "10.38876242041641",
         "13.775937973930814",
         "2.602958679422927",
         "183.59771002934062",
         "22.79878634431914"
        ],
        [
         "('cubic monotonic', 'beta_3', 'parametric_boots')",
         "75.72601216469042",
         "10.38876242041641",
         "0.9841053188719839",
         "2.109768489014601",
         "1828.2970471267927",
         "101.82745503517135"
        ],
        [
         "('cubic unconstrained', 'beta_0', 'Wald/delta')",
         "-75.33868079212766",
         "-197.43456924470524",
         "-162.96251965553296",
         "-402.4329780401599",
         "12.285158071277635",
         "7.56383955074935"
        ],
        [
         "('cubic unconstrained', 'beta_1', 'Wald/delta')",
         "287.5173164621367",
         "248.4905000484067",
         "-51.62426932699947",
         "-14.566758126806093",
         "626.6589022512728",
         "511.54775822361944"
        ],
        [
         "('cubic unconstrained', 'beta_2', 'Wald/delta')",
         "-339.6550928412925",
         "-98.59976324737492",
         "-748.8691888972139",
         "-205.4692297024859",
         "69.55900321462877",
         "8.269703207736043"
        ],
        [
         "('cubic unconstrained', 'beta_3', 'Wald/delta')",
         "124.51434906316405",
         "12.18035307606018",
         "-29.18318823491893",
         "-1.418592629226529",
         "278.211886361247",
         "25.77929878134689"
        ]
       ],
       "shape": {
        "columns": 6,
        "rows": 24
       }
      },
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr:last-of-type th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th colspan=\"2\" halign=\"left\">value</th>\n",
       "      <th colspan=\"2\" halign=\"left\">LCL</th>\n",
       "      <th colspan=\"2\" halign=\"left\">UCL</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th>scale</th>\n",
       "      <th>log</th>\n",
       "      <th>plain</th>\n",
       "      <th>log</th>\n",
       "      <th>plain</th>\n",
       "      <th>log</th>\n",
       "      <th>plain</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model</th>\n",
       "      <th>coef</th>\n",
       "      <th>method</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th rowspan=\"20\" valign=\"top\">cubic monotonic</th>\n",
       "      <th rowspan=\"5\" valign=\"top\">beta_0</th>\n",
       "      <th>Wald/delta</th>\n",
       "      <td>-40.708789</td>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-80.631797</td>\n",
       "      <td>-234.620679</td>\n",
       "      <td>-0.785781</td>\n",
       "      <td>-13.364651</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nonparam_boots</th>\n",
       "      <td>-40.708789</td>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-362.341899</td>\n",
       "      <td>-662.077285</td>\n",
       "      <td>-14.457646</td>\n",
       "      <td>-20.931699</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nonparam_stratified_boots</th>\n",
       "      <td>-40.708789</td>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-369.351128</td>\n",
       "      <td>-647.740990</td>\n",
       "      <td>-14.205809</td>\n",
       "      <td>-21.412504</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>normal</th>\n",
       "      <td>-40.708789</td>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-80.514824</td>\n",
       "      <td>-233.949431</td>\n",
       "      <td>0.576810</td>\n",
       "      <td>-9.562457</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>parametric_boots</th>\n",
       "      <td>-40.708789</td>\n",
       "      <td>-123.992665</td>\n",
       "      <td>-310.971797</td>\n",
       "      <td>-519.918596</td>\n",
       "      <td>-8.945695</td>\n",
       "      <td>-14.267928</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">beta_1</th>\n",
       "      <th>Wald/delta</th>\n",
       "      <td>152.895751</td>\n",
       "      <td>163.794290</td>\n",
       "      <td>-2.372460</td>\n",
       "      <td>14.050524</td>\n",
       "      <td>308.163962</td>\n",
       "      <td>313.538056</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nonparam_boots</th>\n",
       "      <td>152.895751</td>\n",
       "      <td>163.794290</td>\n",
       "      <td>0.008195</td>\n",
       "      <td>7.451974</td>\n",
       "      <td>1941.547918</td>\n",
       "      <td>1082.620980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nonparam_stratified_boots</th>\n",
       "      <td>152.895751</td>\n",
       "      <td>163.794290</td>\n",
       "      <td>0.014464</td>\n",
       "      <td>7.756705</td>\n",
       "      <td>1980.246721</td>\n",
       "      <td>1057.043993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>normal</th>\n",
       "      <td>152.895751</td>\n",
       "      <td>163.794290</td>\n",
       "      <td>34.454749</td>\n",
       "      <td>45.151309</td>\n",
       "      <td>345.981009</td>\n",
       "      <td>347.318752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>parametric_boots</th>\n",
       "      <td>152.895751</td>\n",
       "      <td>163.794290</td>\n",
       "      <td>0.134978</td>\n",
       "      <td>0.269142</td>\n",
       "      <td>1223.357752</td>\n",
       "      <td>858.638061</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">beta_2</th>\n",
       "      <th>Wald/delta</th>\n",
       "      <td>-186.372092</td>\n",
       "      <td>-71.448302</td>\n",
       "      <td>-384.156090</td>\n",
       "      <td>-138.537864</td>\n",
       "      <td>11.411907</td>\n",
       "      <td>-4.358741</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nonparam_boots</th>\n",
       "      <td>-186.372092</td>\n",
       "      <td>-71.448302</td>\n",
       "      <td>-3466.701526</td>\n",
       "      <td>-586.956425</td>\n",
       "      <td>2.548176</td>\n",
       "      <td>-3.339263</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nonparam_stratified_boots</th>\n",
       "      <td>-186.372092</td>\n",
       "      <td>-71.448302</td>\n",
       "      <td>-3528.049144</td>\n",
       "      <td>-576.953166</td>\n",
       "      <td>2.305803</td>\n",
       "      <td>-13.715894</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>normal</th>\n",
       "      <td>-186.372092</td>\n",
       "      <td>-71.448302</td>\n",
       "      <td>-434.565154</td>\n",
       "      <td>-153.877505</td>\n",
       "      <td>-38.043109</td>\n",
       "      <td>-18.651542</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>parametric_boots</th>\n",
       "      <td>-186.372092</td>\n",
       "      <td>-71.448302</td>\n",
       "      <td>-1600.454296</td>\n",
       "      <td>-504.895600</td>\n",
       "      <td>6.086761</td>\n",
       "      <td>-1.304426</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"5\" valign=\"top\">beta_3</th>\n",
       "      <th>Wald/delta</th>\n",
       "      <td>75.726012</td>\n",
       "      <td>10.388762</td>\n",
       "      <td>-8.688065</td>\n",
       "      <td>0.345077</td>\n",
       "      <td>160.140090</td>\n",
       "      <td>20.432448</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nonparam_boots</th>\n",
       "      <td>75.726012</td>\n",
       "      <td>10.388762</td>\n",
       "      <td>2.070028</td>\n",
       "      <td>0.769656</td>\n",
       "      <td>2059.363688</td>\n",
       "      <td>108.791615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>nonparam_stratified_boots</th>\n",
       "      <td>75.726012</td>\n",
       "      <td>10.388762</td>\n",
       "      <td>1.751783</td>\n",
       "      <td>6.602426</td>\n",
       "      <td>2094.675746</td>\n",
       "      <td>106.656112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>normal</th>\n",
       "      <td>75.726012</td>\n",
       "      <td>10.388762</td>\n",
       "      <td>13.775938</td>\n",
       "      <td>2.602959</td>\n",
       "      <td>183.597710</td>\n",
       "      <td>22.798786</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>parametric_boots</th>\n",
       "      <td>75.726012</td>\n",
       "      <td>10.388762</td>\n",
       "      <td>0.984105</td>\n",
       "      <td>2.109768</td>\n",
       "      <td>1828.297047</td>\n",
       "      <td>101.827455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th rowspan=\"4\" valign=\"top\">cubic unconstrained</th>\n",
       "      <th>beta_0</th>\n",
       "      <th>Wald/delta</th>\n",
       "      <td>-75.338681</td>\n",
       "      <td>-197.434569</td>\n",
       "      <td>-162.962520</td>\n",
       "      <td>-402.432978</td>\n",
       "      <td>12.285158</td>\n",
       "      <td>7.563840</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beta_1</th>\n",
       "      <th>Wald/delta</th>\n",
       "      <td>287.517316</td>\n",
       "      <td>248.490500</td>\n",
       "      <td>-51.624269</td>\n",
       "      <td>-14.566758</td>\n",
       "      <td>626.658902</td>\n",
       "      <td>511.547758</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beta_2</th>\n",
       "      <th>Wald/delta</th>\n",
       "      <td>-339.655093</td>\n",
       "      <td>-98.599763</td>\n",
       "      <td>-748.869189</td>\n",
       "      <td>-205.469230</td>\n",
       "      <td>69.559003</td>\n",
       "      <td>8.269703</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>beta_3</th>\n",
       "      <th>Wald/delta</th>\n",
       "      <td>124.514349</td>\n",
       "      <td>12.180353</td>\n",
       "      <td>-29.183188</td>\n",
       "      <td>-1.418593</td>\n",
       "      <td>278.211886</td>\n",
       "      <td>25.779299</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                           value              \\\n",
       "scale                                                        log       plain   \n",
       "model               coef   method                                              \n",
       "cubic monotonic     beta_0 Wald/delta                 -40.708789 -123.992665   \n",
       "                           nonparam_boots             -40.708789 -123.992665   \n",
       "                           nonparam_stratified_boots  -40.708789 -123.992665   \n",
       "                           normal                     -40.708789 -123.992665   \n",
       "                           parametric_boots           -40.708789 -123.992665   \n",
       "                    beta_1 Wald/delta                 152.895751  163.794290   \n",
       "                           nonparam_boots             152.895751  163.794290   \n",
       "                           nonparam_stratified_boots  152.895751  163.794290   \n",
       "                           normal                     152.895751  163.794290   \n",
       "                           parametric_boots           152.895751  163.794290   \n",
       "                    beta_2 Wald/delta                -186.372092  -71.448302   \n",
       "                           nonparam_boots            -186.372092  -71.448302   \n",
       "                           nonparam_stratified_boots -186.372092  -71.448302   \n",
       "                           normal                    -186.372092  -71.448302   \n",
       "                           parametric_boots          -186.372092  -71.448302   \n",
       "                    beta_3 Wald/delta                  75.726012   10.388762   \n",
       "                           nonparam_boots              75.726012   10.388762   \n",
       "                           nonparam_stratified_boots   75.726012   10.388762   \n",
       "                           normal                      75.726012   10.388762   \n",
       "                           parametric_boots            75.726012   10.388762   \n",
       "cubic unconstrained beta_0 Wald/delta                 -75.338681 -197.434569   \n",
       "                    beta_1 Wald/delta                 287.517316  248.490500   \n",
       "                    beta_2 Wald/delta                -339.655093  -98.599763   \n",
       "                    beta_3 Wald/delta                 124.514349   12.180353   \n",
       "\n",
       "                                                              LCL              \\\n",
       "scale                                                         log       plain   \n",
       "model               coef   method                                               \n",
       "cubic monotonic     beta_0 Wald/delta                  -80.631797 -234.620679   \n",
       "                           nonparam_boots             -362.341899 -662.077285   \n",
       "                           nonparam_stratified_boots  -369.351128 -647.740990   \n",
       "                           normal                      -80.514824 -233.949431   \n",
       "                           parametric_boots           -310.971797 -519.918596   \n",
       "                    beta_1 Wald/delta                   -2.372460   14.050524   \n",
       "                           nonparam_boots                0.008195    7.451974   \n",
       "                           nonparam_stratified_boots     0.014464    7.756705   \n",
       "                           normal                       34.454749   45.151309   \n",
       "                           parametric_boots              0.134978    0.269142   \n",
       "                    beta_2 Wald/delta                 -384.156090 -138.537864   \n",
       "                           nonparam_boots            -3466.701526 -586.956425   \n",
       "                           nonparam_stratified_boots -3528.049144 -576.953166   \n",
       "                           normal                     -434.565154 -153.877505   \n",
       "                           parametric_boots          -1600.454296 -504.895600   \n",
       "                    beta_3 Wald/delta                   -8.688065    0.345077   \n",
       "                           nonparam_boots                2.070028    0.769656   \n",
       "                           nonparam_stratified_boots     1.751783    6.602426   \n",
       "                           normal                       13.775938    2.602959   \n",
       "                           parametric_boots              0.984105    2.109768   \n",
       "cubic unconstrained beta_0 Wald/delta                 -162.962520 -402.432978   \n",
       "                    beta_1 Wald/delta                  -51.624269  -14.566758   \n",
       "                    beta_2 Wald/delta                 -748.869189 -205.469230   \n",
       "                    beta_3 Wald/delta                  -29.183188   -1.418593   \n",
       "\n",
       "                                                              UCL               \n",
       "scale                                                         log        plain  \n",
       "model               coef   method                                               \n",
       "cubic monotonic     beta_0 Wald/delta                   -0.785781   -13.364651  \n",
       "                           nonparam_boots              -14.457646   -20.931699  \n",
       "                           nonparam_stratified_boots   -14.205809   -21.412504  \n",
       "                           normal                        0.576810    -9.562457  \n",
       "                           parametric_boots             -8.945695   -14.267928  \n",
       "                    beta_1 Wald/delta                  308.163962   313.538056  \n",
       "                           nonparam_boots             1941.547918  1082.620980  \n",
       "                           nonparam_stratified_boots  1980.246721  1057.043993  \n",
       "                           normal                      345.981009   347.318752  \n",
       "                           parametric_boots           1223.357752   858.638061  \n",
       "                    beta_2 Wald/delta                   11.411907    -4.358741  \n",
       "                           nonparam_boots                2.548176    -3.339263  \n",
       "                           nonparam_stratified_boots     2.305803   -13.715894  \n",
       "                           normal                      -38.043109   -18.651542  \n",
       "                           parametric_boots              6.086761    -1.304426  \n",
       "                    beta_3 Wald/delta                  160.140090    20.432448  \n",
       "                           nonparam_boots             2059.363688   108.791615  \n",
       "                           nonparam_stratified_boots  2094.675746   106.656112  \n",
       "                           normal                      183.597710    22.798786  \n",
       "                           parametric_boots           1828.297047   101.827455  \n",
       "cubic unconstrained beta_0 Wald/delta                   12.285158     7.563840  \n",
       "                    beta_1 Wald/delta                  626.658902   511.547758  \n",
       "                    beta_2 Wald/delta                   69.559003     8.269703  \n",
       "                    beta_3 Wald/delta                  278.211886    25.779299  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Put the two predictor scales next to one another for each model,\n",
    "# coefficient, and uncertainty method.\n",
    "df_beta_CI_comparison = (\n",
    "    df_beta_CI\n",
    "    .set_index([\"model\", \"coef\", \"method\", \"scale\"])[[\"value\", \"LCL\", \"UCL\"]]\n",
    "    .unstack(\"scale\")\n",
    "    .sort_index()\n",
    ")\n",
    "\n",
    "df_beta_CI_comparison\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "a93d455317b9",
   "metadata": {},
   "outputs": [],
   "source": [
    "cubic_monotonic_results = {\n",
    "    \"model\": logit,\n",
    "    \"fit_results\": ress,\n",
    "    \"parameters\": thetas,\n",
    "    \"covariances\": cov_thetas,\n",
    "    \"parameter_ci_wald\": df_theta_CI_wald,\n",
    "    \"parameter_ci_comparison\": df_theta_CI,\n",
    "    \"bootstrap_parameters\": boots_theta_results,\n",
    "    \"theta_parameters\": thetas,\n",
    "    \"theta_covariances\": cov_thetas,\n",
    "    \"theta_parameter_ci_wald\": df_theta_CI_wald,\n",
    "    \"theta_parameter_ci_comparison\": df_theta_CI,\n",
    "    \"bootstrap_theta\": boots_theta_results,\n",
    "    \"beta_parameter_ci_comparison\": df_beta_CI.loc[\n",
    "        df_beta_CI[\"model\"] == \"cubic monotonic\"\n",
    "    ].reset_index(drop=True),\n",
    "}\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e64342b71314",
   "metadata": {},
   "source": [
    "## 5. Collected model results\n",
    "\n",
    "The principal fitted objects and uncertainty results from the three sections are collected below. This makes later cross-model comparisons possible without rerunning or reconstructing section-specific variables.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "4fd30ddf0b43",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['linear', 'cubic_unconstrained', 'cubic_monotonic'])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_results = {\n",
    "    \"linear\": linear_results,\n",
    "    \"cubic_unconstrained\": cubic_unconstrained_results,\n",
    "    \"cubic_monotonic\": cubic_monotonic_results,\n",
    "}\n",
    "\n",
    "model_results.keys()\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base (3.12.3)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}