Identifying Road Accidents Severity Problems Using Data Mining Approaches

Abstract

Roadway traffic safety is a major concern for transportation governing agencies as well as ordinarycitizens. In order to give safe driving suggestions, carefulanalysis of roadway traffic data is critical to find outvariables that are closely related to fatal accidents. Inthis paper we apply statistics analysis and data miningalgorithms on the FARS Fatal Accident dataset as an attempt to address this problem. The relationship betweenfatal rate and other attributes including collision manner,weather, surface condition, light condition, and drunkdriver were investigated. Association rules were discoveredby Apriori algorithm, classification model was built byNaive Bayes classifier, and clusters were formed by simple K-means clustering algorithm. Here we are also using one more classification technique for comparing with Naïve bayes classifier. Certain safety driving suggestions were made based on statistics, association rules, classification model, and clusters obtained.