Context

This is a Python code implementing the MCMC numerical method to estimate parameters of a model.

I developped this code during the COVID pendemic for a project in collaboration with the MODCOV19 CNRS plateform and the ARS of Mayotte. The goal of this project was to estimate the parameters of a COVID propagation model to get more information or compute predictions specificaly for Mayotte. The model simulates the evolution of the number of infected individuals for 8 different age classes. Using probabilities of hospitalization knowing infection for each of these 8 classes, some parameters of the model are estimated to match as closely as possible the observed number of hospitalizations.

This code can be reused to estimate any parameters of a model as long as the user gives a cost function to measure the closeness of the model result to a truth value. In the case of this project, we use the maximum likelyhood of the number of hospitalization for each age class and for each day obtain from the model and compare with the observed data.

Additional information

Related document:

• A report of the ARS with some results given by this computational code: Enquête de séroprévalence à SARS-CoV-2 à Mayotte, 2022.