Mining traffic accident data of N5 national highway in bangladesh employing decision trees

• Published in: 2017 IEEE Region 10 Humanitarian Technology Conference (R10-HTC) pp. 722 - 725
• Main Authors: Satu, Md. Shahriare, Ahamed, Sharif, Hossain, Faruk, Akter, Tania, Farid, Dewan Md
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2017
• Subjects: Accidents; Decision tree; Decision trees; Feature extraction; Feature selection; Roads; Rule extraction; Smart city; Traffic accident data mining; Vegetation; Vehicles
• ISSN: 2572-7621
• DOI: 10.1109/R10-HTC.2017.8289059

Mining traffic accident data is necessary for accident free smart cities, as traffic accidents causes harmful injuries, loss of lives and damages properties of people. N5 National Highway in Bangladesh is the largest highway, where a large number of accidents occur in every year. In this paper, we have analyzed and found the traffic accident patterns of N5 National Highway in Bangladesh using several decision tree induction algorithms. Decision tree is one of the most popular algorithms in machine learning and data mining. We have analyzed total 892 traffic accidents. The traffic accidents data of N5 National Highway of Bangladesh was collected from Modular Accident Analysis Program version 5, Accident Research Institute, Bangladesh University of Engineering and Technology. We have extracted the informative features from the traffic accident data and generated several classifiers using different decision tree algorithms. In the experimental analysis, we have compared the performance of 12 decision tree classifiers and find the best classifier. Finally, we have extracted rules for the trees to avoid traffic accidents in the N5 National Highway of Bangladesh.