Back

1-C Nonlinear Covid-19 Epidemic Model and Application to the Epidemic Prediction in France

Quadrat, J.-P.

2020-05-26 epidemiology
10.1101/2020.05.24.20111807 medRxiv
Show abstract

We have shown in a previous paper that the standard time-invariant SIR model was not effective to predict the 2019-20 coronavirus pandemic propagation. We have proposed a new model predicting z the logarithm of the number of detected-contaminated people. It follows a linear dynamical system[z] = b - az. We show here that we can improve this prediction using a non linear model[z] = b - azr where r is an exponent that we have also to estimate from data. Some countries have an epidemic with a bell shaped form that we call unimodal epidemic. With this new model, we fit observed data of different countries having an unimodal epidemic with a surprising quality. We discuss also the prediction quality obtained with these models at the epidemic start in France. Finally, we evaluate the containment impact on the Covid French mortality in hospitals.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.