关键词: 交通工程；道路安全；心理健康；Logistic回归；主成分分析法；交通事故预测模型

Abstract: A PCA-Logistic traffic accident tendency prediction method based on the psychological status of bus drivers was proposed. Aiming at the problem that some independent variables in the traditional logistic regression model may have multicollinearity, which may affect the prediction results, the principal component analysis method was used to reduce the dimensions of the independent variables, and the PCA-Logistic traffic accident prediction model was constructed. The two classification thresholds were determined by the Jordan index method to represent the occurrence of traffic accidents in the prediction results, so as to achieve the purpose of improvement. Then, combined with the psychological health questionnaire survey results of 1004 bus drivers and the statistical data whether drivers have had traffic accidents in recent one year, the model was verified, and according to the research results, the targeted measures and suggestions to improve the mental health of bus drivers were put forward. The research results show that: among the 9 psychological factors, somatization, obsessive-compulsive symptoms, depression and hostility are the dangerous psychological factors affecting drivers driving safety; the proposed PCA-Logistic traffic accident prediction model is effective, and the threshold division is reasonable.

Key words: traffic engineering; road safety; mental health; Logistic regression; principal component analysis; traffic accident perdition model