This repo is my personal repo for learning how to use bayesian structural time series (bsts) for time series analysis. Using bsts as a foundation, one can perform causal inference of an intervention on time series data by modelling the past and using it as a counterfactual baseline.
Links to readings will be kept on this README document for reference.
Introductory:
- Sorry ARIMA but I'm going Bayesian
- Fitting Bayesian structural time series with the bsts R package
- Structural Time-Series Models
- Spike and slab: Bayesian linear regression with variable selection
- Making Causal Impact Analysis Easy
Package papers and tutorials:
- Predicting the Present with Bayesian Structural Time Series
- INFERRING CAUSAL IMPACT USING BAYESIAN STRUCTURAL TIME-SERIES MODELS
- R package bsts tutorial
- Causal Impact using Bayesian Structural Time-Series Models
- Software for Bayesian Structural Time Series
- MineThatData Forecasting Challenge: proposed solution with Bayesian Structural Time Series models
The following packages may be used in this learning repo:
For my analysis, I used the following data sources: