The best way to conference proceedings by Francis Academic Press

Web of Proceedings - Francis Academic Press
Web of Proceedings - Francis Academic Press

Ensemble learning for insurance premium prediction: a comparative analysis of XGBoost, Random Forest, and SVM

Download as PDF

DOI: 10.25236/ieesasm.2023.066


Danyang Yao, Jiayu Li, Yiqi Shen

Corresponding Author

Danyang Yao


The calculation of premium is an important part for insurance companies to research and introduce new types of insurance. This paper collects the data of insurance companies, and will explore which factors are related to insurance premium from the variables of gender, age, body variable index, whether smoking, number of children and region, and first explore the correlation between factors with correlation coefficient, and then use XGB, random forest and svm model to analyze them one by one. All models show that smoking, age and body mass index have greater influence on insurance premium.


Health Insurace, Random Forest, XGBoost, Support Vector Machines, Forecast