Probabilistic Programming in Python
summary
Demonstrates building a Bayesian logistic regression in Python using Stan and an MCMC sampler on a height/weight/gender dataset, replacing single-point frequentist coefficients with full posterior distributions. Prior distributions are specified for intercept and coefficients, Stan compiles the model and returns posterior samples; summaries show mean/median/std and credible intervals, and predictions integrate parameter uncertainty instead of a point estimate, improving calibrated probability estimates.
AI-written summary · about
chapters (11)
- 0:00intro
- 0:20preamble
- 0:49motivation
- 2:45bayesian vs. frequentist statistics
- 3:57bayes theorem
- 4:46bayesian vs non-bayesian inference
- 6:24bayesian inference
- 7:19markov chain monte carlo (mcmc)
- 8:01probabilistic modelling
- 8:47workflow of probabilistic programming
- 9:43demo
related talks
01h 31mProbabilistic Programming and Bayesian Inference in Python
31mProbabilistic Programming: The What, Why and How
43mProbabilistic Programming in the Real World
01h 30mAn Introduction to Probability and Statistics
01h 26mProbabilistic Programming and Bayesian Computing with PyMC
01h 27mProbabilistic Python: An Introduction to Bayesian Modeling with PyMC
01h 24mBayesianism and Survival Analysis
20mProbabilistic Programming with Stochastic Probabilities