Dengue vector specie’s niche distribution modeling over India using the machine learning-based MaxEnt model

Abstract

NTD (Neglected Tropical diseases), such as dengue, will severely impact public health. So, proper measures and preventive steps should be taken to mitigate dengue outbreaks. This is accomplished by predicting dengue hotspots using SDM models. A well-known Maxent (Maximum Entropy) model was employed to forecast the future spread of vectors based on environmental data, including bio-climatic variables. Accuracy evaluation was performed using AUC values. Historical data on the presence of Aedes aegypti and Aedes albopictus were gathered from GBIF.org (1981–2004), along with corresponding climatic data from (https://chelsa-climate.org/). Features were selected through correlation analysis and AUC optimization, and the model was fitted accordingly. Predictions about future distribution were made under three climatic scenarios, namely SSP126, SSP370, and SSP585, derived from CMIP-6 data. There is a significant aegypti vector distribution over India. Meanwhile, albopictus distribution is less severe compared to the aegypti vector. The vector expansion is visible in all three climatic scenarios, especially in northeastern regions such as West Bengal, partial IGP regions like Madhya Pradesh, and all union territories. The model fitted with utmost accuracy in both training and testing. The aegypti accuracy for training and testing are 0.8081 and 0.7658, and similarly for albopictus, 0.8252 and 0.8056. This analysis will give public health experts a vision for planning mitigation strategies. This was only a preliminary analysis based on environmental modeling rather than mechanistic modeling, which may give more insights. However, climate change will profoundly impact VBD (Vector-Borne Diseases).