Vector Borne Disease Outbreak Prediction by Machine Learning

Abstract

Vector Borne Disease is a form of illness which is caused by parasites, viruses and bacteria. The infection is transferred through blood-feeding arthropods such as mosquitoes, fleas ticks etc. Every year from diseases such as yellow fever, Malaria more than 700,000 deaths occur. These diseases are most common in tropical and subtropical areas and affect the underprivileged populations. Deep learning an essential part of Artificial Intelligence provides an uncanny power to systems to construct a complex network using layers of perceptrons which mimic the human neurons. This network Combined with algorithms of Machine Learning may serve as one of the most powerful tool in healthcare to classify and analyze huge amount of medical data and predict future trends through Supervised Learning. The paper we focused on effective prediction of the vector borne disease outbreak (Multiclass Classification) of three diseases (Chikungunya, Malaria, Dengue) across the Indian-subcontinent. We have examined and refined our model over data collected across India in 2013-2017. We have put forward a Convolutional Neural Network outbreak risk prediction algorithm using contrasting data. To our finest understanding, none of the previous works have centered on contrasting data in area of analysis of medical data. The prediction accuracy of our suggested CNN algorithm is 88%.