Research on Car Insurance Compensation Based on Machine Learning

In the field of insurance claims, it is important to understand the impact of each variable on the amount of the claim. Accurate prediction of whether a car insurance claim will be paid out and the amount of the claim will be paid out is of great significance to increase the company's profit, optimize the premium pricing, reduce the financial risk of the insurance company, and enhance the competitiveness of the insurance company in the market. In this study, we analyze industry reports, market data and relevant literature to study the key factors affecting the amount of insurance claims, firstly, we construct the ‘Insurance Claims Prediction Model’ and select five variables that have the greatest impact on the amount of car insurance claims, and then we predict the amount of car insurance claims through the selected variables. By analyzing the intrinsic connection of the data, based on the random forest model, it predicts whether the insurance is paid or not, and selects the variables that affect the insurance compensation amount, and predicts whether the insurance is paid or not; in addition, in order to accurately predict the amount of the insurance compensation, based on the theory of neural network, it constructs a large-data neural network prediction model, and achieves good experimental results, and obtains a low error. This study conducted the preliminary research on the prediction of car insurance compensation amount, which provides an important reference value for the subsequent formulation of the premium amount of car insurance.