Research on the Extreme Weather Prediction Model Based on Linear Regression and Its Application in Insurance Evaluation

Abstract

This study presents a novel approach to predicting extreme weather events using a linear regression model and explores its application in the insurance evaluation process. The predictive model is developed by analyzing historical data on the frequency and intensity of extreme weather occurrences, such as floods, droughts, and hurricanes. The model employs cumulative data analysis to establish a linear relationship between the occurrence of extreme weather events and the passage of time. The results demonstrate a high correlation coefficient, indicating a reliable prediction capability. The study then extends the application of this model to the insurance sector by constructing an insurance evaluation model. This model assesses the feasibility of providing insurance services in areas prone to extreme weather by comparing insurance premiums with potential claims. The evaluation model incorporates the impact of extreme weather on three major industries: agriculture, manufacturing, and services, and assigns weights accordingly. Case studies of Japan and Florida, USA, are conducted to validate the model's effectiveness in real-world scenarios.