Abstract:
In India, high prevalence of vector borne disease like malaria in certain region is a significant health burden and require extensive community monitoring surveillance and management. Proliferation and spread of all vector borne diseases (including malaria) is known to be tied with several environmental and meteorological factors in addition to human, parasite and vector biology. Modelling and predicting malaria cases is, thus, essential as it can provide an early warning and improve community response in treating the disease. Various time series modelling techniques can be used to model malaria. In this study, we have used Machine learning-based methods like Linear and ridge regression, Self-Organising Maps (SOMs) and LSTM to generate forecast or qualitative district wise outlooks for malaria cases in Bihar. We have used meteorological and health parameters. Our LSTM model for malaria prediction is very flexible and accurate. We have also developed a dashboard for data and model predictions which can later be used as a product for malaria prevention
