Predictive Analysis of Vehicle Insurance Demand Using Machine Learning Techniques

Health insurance and vehicle insurance play a crucial role in reducing economic burdens, alleviating psychological stress, and maintaining social stability. Given the similarities in buyer characteristics for both types of insurance, predicting whether customers who purchase health insurance are interested in vehicle insurance is of significant importance for companies offering both products. By analyzing data from the Kaggle website, this paper employs machine learning techniques, including XGBoost, AdaBoost, and Multi-Layer Perceptron (MLP), to predict customer interest in purchasing vehicle insurance and compares the performance of different models. The study results indicate that AdaBoost performs best in predicting vehicle insurance demand, followed by XGBoost, while MLP performs relatively weaker in this task. These findings provide important insights for insurance companies to optimize marketing strategies, adjust pricing models, and better manage risks. However, future research should focus on integrating more advanced technologies and improving data quality to further enhance the performance of predictive models.