Survival analysis notes

Notes on survival and recurrent event analysis, from several different references.

Contents

1. Generating a smooth estimate of the survival function via the hazard function
2. Recurrent models based on Cox regression (WIP)
3. Using the lifelines library in python to fit KM curves
4. Notes and examples from book "Applied Survival Analysis Using R", by D.F. Moore
5. Confidence intervals for conditional survival estimates
6. Deriving conditional survival distributions based on the Weibull distribution
7. Sampling from arbitrary hazard functions
8. Using the Brier score to evaluate predictive ability of a survival model

Repo structure

• src directory: code files
• .pre-commit-config.yaml: config for use with pre-commit. It specifies what hooks to use. Once this file is created, if you run pre-commit install, the pre-commit tool will populate the pre-commit file in the ./.git/hooks directory. Helpful references:
  • Automate Python workflow using pre-commits: black and flake8
  • Keep your code clean using Black & Pylint & Git Hooks & Pre-commit
  • pre-commit docs
• .flake8: config for Flake8. Mainly used to specify max-line-length=88, to match Black's default
• requirements.txt: python packages used
• renv directory: files created by renv R package to replicate environment. Helpful reference:
  • Introduction to renv

Notes

• I use p2j to convert from .py files to .ipynb files (reference). Unfortunately, this doesn't run the file and create outputs, so I do that manually.