Estimation of parameters of the COVID-19 (SIR model)

This repository allow estimation of parameters of the SIR and SEIR epidemic models.

Dependencies

Firstly, make sure all the dependencies are installed pip install --user --upgrade -r requirements.txt Download the COVID-19 data in the current repository git clone https://github.com/CSSEGISandData/COVID-19

COVID-19 params esitimation

Then the parameters of the COVID-19 can be estimated as python train_random_search.py estimates the parameters on all the countries together, with the assumption of fixed population, equal for each country

To estimate the parameters for a single country (e.g. Italy) python train_random_search.py --train_on_country=Italy --population=60000000 You will see something similar to

The repo allows also to estimate the parameters ot the epidemic models day by day python estimate_params_day_by_day.py --train_on_country=Italy --population=60000000 --train_iters=30000 --predicted_days=60 You will see something similar to

The parameters of the SIRC model on italy, the last --predicted_days are inferred through a linear regression of the last values

The Infectuos / Recovered and Casualties, considering the parameters of the previous plot (note that the parameters vary each day)