Disease Propagation Prediction using Machine Learning for Crowdsourcing Mobile Applications

Being a tropical island, Mauritius is prone to many communicable diseases. Relevant departments of the Ministry of Health and Quality of Life have as mission to monitor and control the propagation of diseases. The primary objective of this paper is to provide the ministry with a system that displays in real time the disease status over the island. The system also predicts the propagation of diseases allowing the ministry to better plan remedial actions. The system consists of three mobile crowdsourcing applications that allow the public, doctors and pharmacies to report diseases and drugs sales in real time. Data regarding diseases for the year 2017 were retrieved and the corresponding daily weather information namely as temperature, humidity and wind for that year was then extracted and added to this dataset. An Artificial Neural Network (ANN) was then trained with this dataset and then used to predict the propagation of the diseases which can be monitored by the Ministry of Health and Quality of Life through another application. The prediction was performed based on the number of reported diseases on the current day along with weather forecasts for the forthcoming days and the results were promising. The model has been evaluated resulting in an accuracy of 90%. Finally, we believe that such a system can be very beneficial to the ministry which can then take informed decisions to counteract the possible propagation of diseases.