Bayesianism and Survival Analysis
summary
Models advertising path-to-conversion with Weibull distribution survival fits, handles heavy censoring, and estimates median lifetimes via MCMC using PyMC2 traces; then fits a Cox model with lifelines to expose covariate effects and computes hazard ratios. Implements Bayesian model averaging with pibma to quantify model uncertainty versus single-model Cox results, shows priors shift posterior medians, and provides runnable notebooks that produce trace plots, median estimates and comparisons that validate the approach.
AI-written summary · about
chapters (9)
- 0:00Bayesian survival analysis intro
- 7:56Data cleaning and censoring
- 21:02First Bayesian model with PyMC
- 35:03MCMC sampling with PyMC
- 44:35MCMC convergence diagnostics
- 53:14Cox proportional hazards model
- 1:07:43Model uncertainty and averaging
- 1:12:07Bayesian model averaging with PIPMA
- 1:23:10Q&A
related talks
21mHazards on the Causal Path
18mCensored - Survival Analysis in Tidymodels
01h 26mProbabilistic Programming and Bayesian Computing with PyMC
34mFast, beautiful and easy Bayesian modelling: Can you have it all?
42mSolving Real-World Business Problems with Bayesian Modeling- 18mCensored - Survival Analysis in Tidymodels
14mProbabilistic Programming in Python
01h 26mNew Trends in Estimation and Inference