Research on Prediction of Anti-Fraud in Automobile Finance Based on XGBoost Machine Learning Algorithm
Yong Bao, Haibiao Wen · 2024
XGBoost, a machine learning model focused on a gradient lifting algorithm, has a good learning effect and fast computing speed; therefore, it is one of the best integrated technologies. Currently, XGBoost is widely used for this purpose. In recent years, the application of machine learning algorithms in financial anti-fraud research has become increasingly popular, but it mostly focuses on credit card transactions, online loans, real estate transactions, etc. However, there is little research on auto finance anti-fraud. In this study, the XGBoost algorithm model, which has been one of the most successful machine learning models in recent years, is used for prediction research based on the sample data of auto loans of Internet finance companies. In the XGBoost algorithm model, the XGBoost algorithm with default parameters and the XGBoost algorithm with optimized parameters were used for modeling and prediction. Finally, the anti-fraud of automobile finance prediction is compared and analyzed using the XGBoost learning algorithm and logical regression model, which are widely recognized in the industrial field. It is pointed out that the XGBoost machine learning model performs better on “AUC, Accuray value and recall rate (recall).” This is because the XGBoost machine-learning model includes more dimensional information, a more scientific way to deal with missing values, and the algorithm principle of considering regularization terms.