BayesBlend real-world examples

This repository contains the scripts used to reproduce the analyses in "BayesBlend: Easy Model Blending using Pseudo-Bayesian Model Averaging, Stacking and Hierarchical Stacking in Python" by Haines & Goold (2024).

Install requirements

All analyses were run with Python 3.11. Once Python 3.11 is installed locally, we recommend the following steps:

1. navigate to your local stacking-paper-2024/ directory
2. initialize a virtual environment: python3.11 venv env
3. activate the environment: source env/bin/activate
4. install requirements: pip install -r requirements/requirements.txt

Reproduce analyses

Analyses can then be reproduced by running the following scripts in order:

1. download and pre-process the data: python -m data-prep
2. fit the loss development models and produce figures: python -m development
3. fit the loss forecasting models and produce figures: python -m forecast

Once analyses are reproduced, figures are located in the stacking-paper-2024/figures/ directory.