Predictive spatial correlation analysis of snakebites of Krishna District, India

Every year, many farmers die as a result of snakebites or other envenomations all over the world. There is currently a need to identify the snakebite regions that cause severe damage in the Indian farmer’s family, especially in the study area. The study area, i.e., India’s Krishna district, is highly prone to snakebites and floods. It best suits the high yield of agricultural crops due to fertile lands and rich water resources. The present research focuses on the identification of the snakebites-prone areas using geospatial analytics based on the Area Land (AL), Rainfall (RF), Flood Zone (FZ), Snakebite (SB), River Side (RS), and Locality (L). A novel Snake Bite Impacted Area (SBIA) method is proposed and implemented to find the Geospatial predictive correlation analysis of snakebites. To measure the performance of SBIA, various popular data classification methods are tested, such as multi-class SVM, Decision Trees (DT), Boosted Trees (BT), Linear Discriminate (LD), and K-Nearest Neighbor (KNN). Among all popular techniques, the SBIA performance is high with an accuracy of 94.12, whereas the accuracy of LD is 76.5, then both SVM and KNN are 82.4, whereas DT and BT are 94.1. In addition the AUC result of SBIA is compared with existing Ecological Ninche Models to predict snakebite prone areas with location. Overall, the performance of SBIA is good, which combines spatial and non-spatial data to predict the snakebite-prone areas.