Research on Quantitative Method of Traffic Safety Credit Score Based on Ridge-Logistic Regression

In order to quantitatively analyze the driver’s traffic credit score, a quantitative analysis method based on Ridge-Logistic regression was proposed from four aspects: selecting drivers’ characteristics, defining good and bad individuals, determining the weight coefficients of different values of each characteristic, and improving the interpretability of the algorithm output results. And the output result of the algorithm was converted into a standard score table form through a score formula. The research results show that the model is highly interpretable, and the score table results are generally in good condition. According to the data test, the accuracy, precision, recall and Area under Curve (AUC) of the model are 98.31%, 97.36%, 99.33% and 0.99, respectively, which means that the model can correctly classify bad individuals and also has a good recognition effect on good individuals. The research results can be used to quantify the actual traffic credit and help to establish a reasonable traffic credit scoring mechanism.