Lay abstract for Applied Stochastic Models in Business and Industry article: Automobile Insurance Claim Occurrence Prediction Model Based on Ensemble Learning

Each week, we publish lay abstracts of new articles from our prestigious portfolio of journals in statistics. The aim is to highlight the latest research to a broader audience in an accessible format.
The article featured today is from Applied Stochastic Models in Business and Industry with the full article now available to read here (OPEN ACCESS).
Si, JHe, HZhang, JCao, XAutomobile insurance claim occurrence prediction model based on ensemble learningAppl Stochastic Models Bus Ind20221– 14. doi:10.1002/asmb.2717
Automobile insurance occupies an important position in property insurance, and in recent years, the wave of the information age has brought automobile insurance into the era of big data. In this context, the generalized linear model (GLM) used by traditional actuaries cannot provide an accurate portrayal of the data. How to find more suitable tools to improve the accuracy of risk loss prediction, more reasonable rate setting and develop automobile insurance products to meet the new era has become an urgent issue for insurance companies. In order to solve the above problems, this paper constructed a two-layer Stacking model and introduced it to the automobile insurance domain. The article first balanced the categories of the three datasets using the Synthetic Minority Oversampling Technique (SMOTE). Then four algorithms including Logistic Regression, Random Forest, Gradient Boosting Decision Tree (GBDT) and Two-Layer Stacking Model, were used to model the occurrence of auto insurance claims respectively. By comparing the AUC values and F1-score, it was found that the constructed model improved in terms of prediction accuracy compared to the other three algorithms. The article also concluded by using the predicted probabilities to develop transfer rules to obtain a more reasonable bonus-malus system. The results showed that the algorithm proposed in this paper can greatly improve the accuracy of loss prediction, and the transfer rule developed can make rate determination more reasonable, which is of great significance to promote the healthy development of insurance companies. 
More Details