Flood Risk Assessment and Interpretation Using Explainable AI

Floods are the unforeseen situations to the society which causes high risk to the society. For proactive action, it is important to identify locations based on past results. It is necessary to take measures before the situation gets worst. For the benefit of society, a novel approach to map flood risk usinga balanced dataset and an Extra Tree Classifier. The Extra Tree Classifier performs better with the Shapely Additive explanation values. It shows improvement towards decision making processes to reduce the causality at the time of floods. The proposed method achieves the accuracy of 93.61% for the estimating flood danger effects on a particular location. In this paper quantitative study was done to handle the challenges to predict the flood danger using the artificial intelligence ethically.