부산대학교 도서관

In order to ensure the security of power intelligent terminals, an identity identification method based on behavior feature modeling was proposed. Firstly, user behavior data was collected. Then the user behavior was modeled according to the collected data. Finally, XGBoost regression model was used to identify and classify the user identity. The results showed that the identification accuracy, recall rate and F1 value of user identity identification through behavior modeling and XGBoost regression reach more than 90%, which has a good identification effect; Compared with SVM algorithm and random forest algorithm, the XGBoost regression has better identification effect, and its identification accuracy is 93.09%. Therefore, the proposed identity identification method based on behavior modeling and XGBoost regression can more accurately identify the user identity, so as to improve the security of using terminal.