Explainable Machine Learning Models for Predicting COVID-19 Cases in Catalonia Based on Wastewater Monitoring Data

Surveillance of infectious diseases that can spread quickly among the population has become a critical goal nowadays due to the dramatic effect of diseases like SARS-CoV-2. One promising method to be able to monitor the spread of such diseases among the population of a city is the analysis of biological compounds in the sewage network of different cities. In this paper, we summarize the results of training a prediction model for SARS-CoV-2 cases based on historical biological data collected from different wastewater treatment plants (WWTP) located in different parts of Catalunya. We consider different approaches for the prediction problem and develop models based on extreme gradient boosting. Finally, we evaluate the quality of the results of our model and study the relevance of the different variables of the model using SHapley Additive exPlanations (SHAP) analysis.