A driver abnormal behavior warning method based on isolated forest algorithm

Accurate monitoring of abnormal driving behavior can effectively improve the safety and efficiency of drivers, reduce accident rates, and improve the traffic environment, which has become a key research topic in the field of intelligent transportation. To improve the accuracy of monitoring abnormal driving behavior, this paper designs a driver abnormal behavior warning method based on the isolated forest algorithm. Firstly, by analyzing abnormal driving behavior, the XGBoost algorithm is used to extract abnormal driving behavior features; Secondly, by constructing an isolated forest of abnormal behavior, the abnormal behavior is divided into single and complex anomalies, and a detection model for abnormal driving behavior is established; Finally, the entropy weight method is used to determine the weight of behavior combinations and calculate the threshold for abnormal behavior scores to achieve early warning. The experimental results show that the accuracy of abnormal driving behavior detection using this method can reach 98.6%, with a time consumption of only 1.3 seconds and a recall rate of 98.6%, indicating that this method can improve the effectiveness of behavior warning.