Bayesian Network of Traffic Accidents in Malaysia

Abstract

Exploring the causes and effects of a hazardous event such as traffic accidents have been of vital importance to society. Statistical analyses have been widely implemented to understand and deduce inferences on the cause-effect analysis, and to anticipate the occurrences of accidents in the future. One of the issues that has not been solved through conventional statistical modelling is the existence of interrelationships between variables in the data set. However, with the advent of technology and the wide application of machine learning algorithm, this problem can be solved through the application of Bayesian network analysis, which is a directed acyclic probabilistic graphical model. By using Hill Climb (HC) and Tabu algorithms, the structure of the data was studied and the relationship was estimated through conditional probability, that is based on the Bayes' theorem. The results suggests that weather plays a major role in the increase of traffic accidents, and occurs by disrupting lighting conditions which then disrupts the traffic systems. Furthermore, the results indicate that fatal accidents have a higher likelihood to occur in head-on, turn over and out of control accidents. The use of the Bayesian network creates probability estimates to enable the identification of the risk and the necessary precaution needed to be implemented.

Towards Sustainable Road Safety in Saudi Arabia: Exploring Traffic Accident Causes Associated with Driving Behavior Using a Bayesian Belief Network

“…BBN has been applied to analyze traffic accidents in different countries. For example, Zou and Yue [47] studied accident causation in Australia, Karimnezhad and Moradi [48] studied the causes of accidents in Iran, Deublein et al [49] studied the causes of accidents in Switzerland, and Zamzuri et al [50] studied the causes of accidents in Malaysia. However, to the authors' knowledge, no BBN model has been proposed for road safety modeling in Saudi Arabia.…”

Analysis of the causes of pedestrian-vehicle traffic accidents based on Bayesian networks

“…For example, Keay and Simmonds 17 designed a regression model for the effect of weather on tra c; Yu 18 established a highway collision model using Poisson model and Bayesian inference method. Selvaso a and Arulraj 19 used GIS for tra c accident analysis; Zamzuri 20 used Bayesian network and HC algorithm and Tabu algorithm to explore the causes of tra c accidents in Malaysia; Zhang 21 used text processing technology based on LDA Topic Model to analyze tra c accidents to obtain the most dominant factors of tra c accidents; Deng 22 proposed a causality analysis model for tra c accidents using a hybrid AHP and Apriori-Gentic based algorithm to mine accident causes. Bayesian network is a good way to model causality, and Bayesian network can effectively form probabilistic models for e cient inference and learning 23 .…”

Spatio-Temporal Clustering of Road Accidents in Kelantan, Malaysia

“…A study done by Zamzuri et al (2019) identified accident variables such as weather, light condition, traffic system, lane marking, road geometry, collision type and accident severity when investigating the interrelationship between these variables to further understand the cause of accidents. Furthermore, many studies reported on the causes or factors that are closely related to accident severity specifically on accidents occurrence and their locations.…”