Data-driven typhoon case retrieval method based on spatiotemporal trajectory similarity and risk preferences of decision-makers

Abstract

The results of typhoon case retrieval are important for decision-making during emergency situations. Therefore, this study proposes a new typhoon case retrieval method to improve the accuracy of the retrieval results based on the spatiotemporal trajectory similarity and the risk preferences of decision-makers. First, a spatiotemporal trajectory similarity measurement is proposed to generate the path similarity case set. Next, attribute similarity and case similarity measurements are applied to obtain the set of similar cases. Subsequently, a group decision-making model based on fuzzy evidence reasoning is constructed to obtain the risk utility value according to the assessment information and risk preferences of the decision-makers. The risk utility and case similarity are aggregated to obtain the comprehensive utility value and the most appropriate historical typhoon case, thereby generating the emergency alternative. Finally, the effectiveness and feasibility of the proposed method are demonstrated using a typhoon disaster example. The advantages of the proposed method are illustrated through comparative analyses with other methods.