qra/organized_uncertainty_analysis

Organized uncertainty analysis

This folder is a self-contained, model-first organization of the thesis code. The original files in FINAL FILES are preserved unchanged.

Install the reproducible environment with python -m pip install -r requirements.txt.

Structure

organized_uncertainty_analysis/
├── data.py
├── suv_percentilesSLOthenUWM.mat
├── flags_combined.mat
├── logistic/
│   ├── core.py
│   ├── data_fit_gof.py
│   ├── ci_estimation.py
│   ├── elasticity.py
│   ├── noise_measurement.py
│   └── notebooks/
│       ├── 01_data_fit_gof.ipynb
│       ├── 02_ci_estimation.ipynb
│       ├── 03_elasticity.ipynb
│       └── 04_noise_measurement.ipynb
├── bayesian/
│   ├── core.py
│   ├── noise_core.py
│   ├── data_fit_gof.py
│   ├── ci_estimation.py
│   ├── elasticity.py
│   ├── noise_measurement.py
│   └── notebooks/
│       ├── 01_data_fit_gof.ipynb
│       ├── 02_ci_estimation.ipynb
│       ├── 03_elasticity.ipynb
│       └── 04_noise_measurement.ipynb
└── outputs/

Recommended order

Run notebooks 01 through 04 within either model folder. Each notebook is self-contained and locates the project root automatically. Expensive replication counts are defined near the top of the relevant notebook so test and thesis runs can be distinguished clearly.

To reproduce all committed tables and figures without Jupyter, run python run_all_notebooks.py from the project root. Generated PNG/PDF figures and CSV tables are written to outputs/. Keep these outputs under version control when they are part of the thesis results.

Module responsibilities

• data_fit_gof.py: data loading, FULL/TRIM construction, fitting, prediction, and goodness-of-fit.
• ci_estimation.py: Wald, MCA, NPBS, PBS, parameter intervals, risk bands, and intervals for derived characteristics.
• elasticity.py: local elasticity of x50 and s50.
• noise_measurement.py: additive and multiplicative biomarker-noise propagation.
• core.py: preserved underlying implementation used by the four focused public modules.

All generated tables and figures should be written to the top-level outputs/ directory.