Build Facebook's Prophet in PyMC3
summary
Implements Facebook Prophet’s model inside PyMC3 reproducing the linear growth with 25 change points grid, vectorized dot‑product offsets and the seasonal components as Fourier series seasonality; uses a Laplace prior for sparsity on changepoint adjustments and runs PyMC3 MCMC sampling to draw posterior samples. Shows posterior histograms, credible intervals and decomposed trend/yearly/weekly components, reports ~30 minutes sampling live and produces forecasts with visible uncertainty that match the Prophet-style behavior.
AI-written summary · about
chapters (8)
- 0:00Prophet and PyMC3 intro
- 3:31Bayesian modeling basics
- 5:00Bayesian inference with PyMC3
- 9:13Trend modeling with change points
- 16:35Adding seasonality with Fourier
- 20:40Complete Prophet model in PyMC3
- 23:10Interpreting forecast results
- 27:09Summary
related talks
42mSeanTaylorBenLetham
41mNeural Prophet – A powerful AI framework for Time Series Models
01h 26mProbabilistic Programming and Bayesian Computing with PyMC
01h 27mBayesian Time Series
42mSolving Real-World Business Problems with Bayesian Modeling
44mAn introduction to PyMC3
35mWhat we learned by running a large custom Bayesian...
34mFast, beautiful and easy Bayesian modelling: Can you have it all?