Investigating the Casual Effect in Traffic Accident

Abstract

An important field of study that attempts to increase road safety and lower the frequency and severity of accidents is the investigation of traffic accidents. For the purpose of creating effective preventative methods and policies, it is imperative to comprehend the underlying causes of traffic accidents. The practice of analyzing the relationship between two or more variables to ascertain whether one has a causal effect on the other is known as causal analysis. Through the integration of mutual information for causality analysis and Support Vector Machine (SVM) for prediction, this system is intended to examine the causative impacts of traffic accidents. The system primarily looks at the reasons behind traffic accidents in Thailand between 2016 and 2019, trying to pinpoint important elements and create practical preventative measures. The system gathers a wealth of information, such as the date, time, and position of each collision as well as information on the type of vehicle, the characteristics of the road, driver demographics, and weather. Mutual information is used to quantify dependencies, highlight important interactions, and study the causal linkages between various components. These analyses show how changes in one variable may have an impact on another. By concentrating on the most important variables, the mutual information and SVM integration improves the system's analytical skills and improves model accuracy and interpretability. As a result, our technology produces thorough reports and visualizations that give stakeholders—such as legislators and traffic safety authorities—actionable insights. These observations aid in the creation of focused initiatives and laws meant to lower the frequency and seriousness of traffic accidents in Thailand.