基于随机森林与逻辑回归模型的交通事故严重程度的预测研究

Prediction of Traffic Accident Severity Based on Random Forest and Logistic Regression Model

摘要: 交通安全与人们的生活紧密相关，交通事故造成的严重程度对社会和人们的生活都有极大的影响，本文选取随机森林与逻辑回归算法构建了交通事故严重程度预测模型，对交通事故严重程度进行了预测与对比分析，对比分析显示出随机森林模型有更好的预测效果，并将影响交通事故严重程度的特征进行重要性排序，可以判断哪些因素对交通事故严重程度影响较大，为交通道路基础建设，以及交通事故严重性预防和降低提供了参考与建议。

Abstract: Traffic safety is closely related to people’s lives. The severity of traffic accidents has a great impact on society and people’s lives. This paper chooses random forest and logistic regression algorithm to construct a traffic accident severity prediction model, and makes a prediction and comparative analysis of the severity of traffic accidents. It shows that stochastic forest model has better prediction effect, and ranks the characteristics that affect the severity of traffic accidents. It can judge which factors have greater impact on the severity of traffic accidents. It provides reference and suggestions for traffic road infrastructure construction, as well as for the prevention and reduction of the severity of traffic accidents.