Hierarchical Bayesian Modelling with PyMC3 and PySTAN
summary
Compares PyMC3 vs PySTAN on a real dataset of ~2500 cars and the Volkswagen case, implementing hierarchical linear regressions with manufacturer and parent-company levels, showing hierarchical partial pooling effects, priors used (Student T, half‑Cauchy), MCMC settings (NUTS, 4 chains), timings (Stan compiles faster, PyMC3 faster prototyping), code and data on GitHub, and ends with Volkswagen mid‑pack rather than a smoking gun.
AI-written summary · about
chapters (8)
- 0:00Bayesian modeling tools intro
- 3:41PyMC3 vs PyStan
- 9:28Car emissions dataset
- 16:40Pooled linear regression
- 25:25Unpooled and hierarchical models
- 30:28Partial pooling with manufacturers
- 33:41Model performance insights
- 35:54Q&A
related talks
34mFast, beautiful and easy Bayesian modelling: Can you have it all?
44mAn introduction to PyMC3
42mBayesian Non-parametric Models for Data Science using PyMC3
42mSolving Real-World Business Problems with Bayesian Modeling
36mClusterf*ck: A Practical Guide to Bayesian Hierarchical Modeling in PyMC3
01h 26mProbabilistic Programming and Bayesian Computing with PyMC
57mA Quickstart Guide to PyMC3
01h 30mAn introduction to Markov Chain Monte Carlo using PyMC3