Providing Intelligent Models for Traffic Accident Data Classification based on Bayesian Network, Decision Tree and Artificial Neural Network

Document Type : Original Article

Authors
َAli Naderan 1
Mohammad Reza Ahadi 2
Rasoul Zabihian 3
1 Assistant Professor, Department of Civil Engineering, SR.C., Islamic Azad University, Tehran, Iran.
2 Associate Professor, Transportation Research Institute, Road, Housing and Urban Development Research Center, Tehran, Iran.
3 Ph.D., Candidate, Department of Civil Engineering, Kish International Branch, Islamic Azad University, Kish Island, Iran.
10.22034/road.2025.498744.2357
Abstract
Background and purpose: Every year, a large number of road users die or are injured as a result of traffic accidents. Analyzing accident-prone points for driving, transportation and traffic organizations, municipalities and organizations related to traffic safety is very important. In order to identify high-risk points of accidents, methods such as nearest neighbor are used, but these methods have insufficient prediction and incorrect estimation percentage. The purpose of this research is to classify accident data and provide a better method for identifying high-risk points in terms of accidents.

Research method: For this purpose, the performance of decision trees, Bayesian model, and neural network were compared using common methods such as nearest neighbor and fitting.Using GIS tools and density-based methods such as kernel density function, high-risk points of accidents were identified in the form of functional maps.

Findings: The results showed that decision tree method, Bayesian network and neural network have better performance than traditional methods with 83%, 83% and 78% percentage of correct estimation, respectively. Also, the results showed that the number of accidents in peak traffic periods, for example morning and evening, is higher than non-peak periods.

Conclusion: By examining the accident-prone points after identification, it can be concluded that the inner-city accident-prone points are mostly near intersections and often occur in suitable weather conditions and in sufficient daylight
Keywords
Urban Accidents
Classification
GIS
Bayesian Network
Decision Tree and Artificial Neural Network
-Elvik, R., (2008). A survey of operational definitions of hazardous road locations in some European countries. Accident Analysis & Prevention, 40(6).1830-1835.
-Kemal Polat, S.S.D. Subtractive clustering attribute weighting (SCAW) to discriminate the traffic accidents on Konya–Afyonkarahisar highway in Turkey with the help of GIS: A case study. Advances in Engineering Software, 2011. 42(7). 491-500.
-V.Prasannakumar, H.V. R.Charutha, N.Geetha, (2011). Spatio-Temporal Clustering of Road Accidents: GIS Based Analysis and Assessment. Procedia - Social and Behavioral Sciences, 21, 317-325.
-Jafari M, Tabrizi A, Esmaeilzadeh MH, Zamani A. (2022). Determining of Accidental Points of Motorcycle Accidents Using Geographic Information System (GIS) in Gonabad and Bajestan Counties, North-East Iran. Irtiqa Imini Pishgiri Masdumiyat. 10(3), 262-268.
-Romi Satria, M.C. (2016). GIS Tools for Analyzing Accidents and Road Design.
A Review. Transportation Research Procedia, 18. 242-247.
 
 
-Saffet Erdogan, I.Y. (2008). Tamer Baybura, Mevlut Gullu, eographical information systems aided traffic accident analysis system case study: city of Afyonkarahisar. Accident Analysis & Prevention, 40(1),174-181.
-SavaşDurduran, S. A (2010). decision making system to automatic recognize of traffic accidents on the basis of a GIS platform. Expert Systems with Applications, 37(12). 7729-7736.
-Al-Ghamdi, A.S. (2003). Analysis of traffic accidents at urban intersections in Riyadh. 717–724.
-József Benedek, S.M.C. (2016). Titus Cristian Man, Hotspots and social background of urban traffic crashes: A case study in Cluj-Napoca (Romania). Accident Analysis & Prevention, 87. 117-126.
-Ali Soltani, S.A. (2017). Exploring Spatial Autocorrelation of Traffic Crashes based on Severity, in Injury. 637-647.
-Seiji Hashimoto, S. RyokoSaeki, YasuhiroMimura, RyosukeAndo, ShutaroNanba, (2016). Development and application of traffic accident density estimation models using kernel density estimation. Journal of Traffic and Transportation Engineering, 3(3). 262-270.