X.J. Zhao, S.Y. Li, H.B. Guan

Pages: 39-50

Abstract
The traditional methods for analyzing environmental impact factors in urban road traffic accidents suffer from issues such as low recall, poor precision, and time-consuming processes. To address these challenges, a novel analysis method based on the logistic regression model is proposed. This method utilizes the MIFP-Apriori algorithm to collect urban road traffic accident data and incorporates the EM algorithm for data filling and processing. Principal component analysis is then employed to screen the environmental impact factors in urban road traffic accidents. These selected factors are input into the logistic regression model to conduct a comprehensive analysis. The results reveal that environmental impact factors like road conditions and meteorological factors have a significant influence on traffic accidents. Experimental results show that the proposed method achieves an average recall rate of 96.98%, an average precision rate of 97.66%, and the analysis time for environmental factors ranges from 0.12s to 0.52s.
Keywords: logistic regression model; urban road; traffic accident; environmental impact factors; mifp-apriori algorithm; principal component analysis method