A study of the impact of extreme weather events on property insurance based on ARIMA time series model and SVM classifier

Abstract

The thesis aims to tackle the challenges posed by the frequency of extreme weather events globally to the property insurance industry. The essay effectively predicts the frequency of extreme weather events and their economic impact on the insurance industry by developing an integrated analytical framework that combines insurance and coverage models. Firstly, an ARIMA time series model was used to forecast future extreme weather events and combined with the Spearman correlation coefficient (SCC) to quantify the relationship between insurance company revenues and socio-economic factors in the insured region, which were used as inputs to a support vector machine (SVM) classification model to assess risk and determine the amount of insurance coverage. In addition, the variables affecting decision-making were adjusted by the entropy weight method to further enhance the accuracy and usefulness of the model. The results of the study show that the constructed model has a high accuracy of 95% in predicting extreme weather events and their impact on the insurance industry, providing a powerful risk management and resource allocation tool for insurance companies, and helping to improve the resilience of property and the overall resilience of the insurance industry.