Toward a Data Fusion Based Framework to Predict Schistosomiasis Infection

We propose a conceptual framework to predict the risk of freshwater source infestation by Schistosomiasis parasites. Our approach aims to combine two sources of information which are outputs of prediction models. The proposed framework is broken down into three Y-shaped branches. The left branch is a water quality prediction model built on the basis of machine learning algorithms applied on data collected by an IoT platform. These data represent physical and chemical parameters of a freshwater source which affect the development of snails and parasites that cause Schistosomiasis. The branch on the right is a non autonomous mathematical model which through its derived reproduction number $R_{0}$ determines the density evolution of all actors involved in Schistosomiasis transmission life cycle. In the middle branch happens a fusion process which combines the two information by taking into account their uncertainty and complementary. The output of the fusion is the final decision about the risk of infestation. This work has focused on the identification of applicable machine learning algorithms for water quality prediction and the identification of a mathematical model. The work has consisted also to give the characteristics of the fusion problem to handle.