Long-term Impact Assessment of Disasters through Predictive Analytics

Disaster is a significant problem that extensively affects society and the community. Predicting the effects of a disaster is difficult for several reasons. The primary aim of this study is to evaluate the effects of disasters across several timeframes, ranging from immediate to long-term. To construct a plausible model, the proposed solution considers the available disaster datasets from various agencies (e.g. SMS, ISC, NDMC etc.). A methodology for assessing the long-term effects of disasters utilizing well-liked machine learning techniques is presented here. It consists of the algorithms for Decision Tree, Random Forest, Gradient Boost Decision Tree and XG Boost. The algorithms' classification accuracy for the provided data sets is 56%, 63%, 83% and 91% respectively. The proposed work also examines the various levels of disaster severity and suggests solutions for each level to improve preparedness and response measures.