Y.W. Chen, P. You, Z.K. Chang

Pages: 13-26

Abstract
Urban road traffic safety is of great significance as it directly affects the well-being of citizens and the efficient operation of cities. To precisely assess the influence of diverse factors on urban road traffic safety, this article employs binary logistic regression analysis. The factors related to urban road traffic safety are categorized into four aspects: human, vehicle, road, and environment. By using the occurrence of a traffic accident as the dependent variable and calculating the odds ratio to gauge the impact of independent variables, after conducting multicollinearity tests on the binary logistic regression model, the regression coefficient analysis reveals that driver behavior, weather conditions, road conditions, and lighting conditions have a substantial impact on urban road traffic safety. Experiments demonstrate that the coverage rate of influencing factors in this paper exceeds 90%, with a comprehensiveness of up to 95%. This indicates that the binary logistic model utilized in this study for factor regression analysis of urban road traffic safety exhibits high coverage, comprehensiveness, a broad regression scope, and high reliability and robustness.
Keywords: traffic safety; binary logistic regression; multicollinearity test; regression coefficient; normalization processing