Logistic Model for Predicting Traffic Fines: A Statistical Approach to Demographic and Behavioral Factors

Abstract

This study uses a logistic regression model to analyze the factors influencing the likelihood of receiving traffic fines within a year. The dichotomous dependent variable (yes/no) was modeled based on independent variables such as age, previous accidents, rest and vacation, active breaks, and general intensity. The results indicate that lack of rest, previous accidents, and absence of active breaks significantly increase the likelihood of receiving fines, while age and general intensity act as protective factors. The model demonstrated robust fit (AUC = 0.9101) and a Pseudo R-squared between 45.6% and 61.2%, showcasing its predictive capability. These findings can inform strategies to reduce traffic violations through interventions in driving habits and workplace wellness.