Prediction model of insurance profit and loss under extreme weather based on PCA and ARIMA Models

Abstract

Because global catastrophe insurance gap widening, insurance profit forecast is important for ensuring the operation of insurance companies under extreme weather. In this paper, on the basis of analyzing the profit structure of insurance, the dimensionality reduction capability of PCA and the flexibility of ARIMA are combined to establish a prediction model of insurance profit and loss under the extreme weather. On the basis of using PCA (Principal Component Analysis) to extract the important features of extreme weather data and to conduct correlation analysis, the influences of each factor on the expression of different principal components are clarified, and the calculation formula of the principal component is obtained. ARIMA (Autoregressive Integrated Moving Average Model) is used to fit historical catastrophe insurance data and capture trends, to predict the likely future compensation ratio in this region. This method can not only reduce the dimension of complex weather data to simplify the analysis, but also flexibly predict the risk degree of future losses of insurance companies. The results show that with the increase of the frequency of extreme weather and the maximum rainfall, the operating risk of insurance companies will rise. Finally, in order to better develop regional catastrophe insurance, it’s a good choice to carry out long-term and short-term planning according to the prediction model.