Forecasting cases of RSV using artificial neural networks and mechanistic models

We study and present an approach based on artificial neural networks to forecast the number of cases with the Respiratory Syncytial Virus (RSV). The number of cases of RSV in most of the countries around the world present a seasonal type behavior. We construct and develop several multilayer perceptron models that intend to forecast appropriately the number of cases of RSV. We compared our approach with a classical technique for time series, and our results are more accurate. The adjusted MLP network that we find has a fairly high accuracy of forecast. Finally, we compare empirical and mechanistic models applied to forecasting and prediction.