Malaria Incidence Prediction Using Leaky Integrate and Fire Neuron

Abstract

A malaria parasite is transmitted to humans with female anopheles bites. Malaria is a common disease in rural and urban areas of most countries. Malaria affects public health and flourishes mainly in sub-tropical countries and tropical countries. The impact of the disease is high, whereas the prevention control system facilities are limited. A prominent prediction model is needed to overcome the effects of malaria disease and prevent infected humans from it. In this study, we focus on determining the malaria-abundant regions with the help of environmental factors and utilizing spiking neural networks as a classification/prediction model for the Ponda region in Goa, India. Groundwork findings of malaria-abundant areas are collected using clinical data from a renowned regional hospital. Leaky integrate and fire neuron models exhibited interesting spiking patterns, which are further used for classifying the malaria-prone zones. This model also demonstrated the best prediction accuracy of 98.47% and 90.87% for village Boma using the MalariaDataset2 and Curti using the MalariaDataset1, respectively. This study tries to contribute a step forward in making India a malaria-free country, a target set by World Health Organization. © 2023 IEEE.