/covid19-backtesting-publication

Primary LanguageJupyter NotebookApache License 2.0Apache-2.0

Backtesting the predictability of COVID-19

This repository contains the relevant code to replicate the main results of our paper "Backtesting the predictability of COVID-19" found at https://arxiv.org/abs/2007.11411.

Running backtesting

To run all historical fits of SEIRD and Power Growth models, run

python fit.py

Generating figures

To generate the figures of Section 4 Experiments of the article, run the paper_plots.ipynb notebook accordingly. It uses the stored backtested parameters from the respective json files in the files folder.