Abstract: Road traffic violations are one of the major causes of road accidents worldwide. Traditional traffic monitoring systems rely heavily on manual enforcement and reactive measures, which are inefficient in reducing violations proactively. This research proposes a Machine Learning-based Traffic Violation Prediction System that predicts the probability of traffic violations based on various input parameter such as driver behaviours, vehicle characteristics, location type, and time conditions. The proposed system utilizes classification algorithms to determine the likelihood of violation occurrence and provides multi-dimensional output including risk level, probability score, and safety recommendations. Experimental results show improved prediction accuracy and decision-support capability. The system can assist traffic authorities in proactive enforcement and smart city development.

Keywords: Machine Learning, Traffic Violation Prediction, Road Safety, Classification, Smart Traffic System, Risk Analysis

Downloads: | DOI: 10.17148/IARJSET.2026.13256