Risk analysis of road traffic accidents based on improved data mining method. (11th January 2023)

• Record Type: Journal Article
• Authors: Feng, Tianjun; Gao, Tan
• Is Part Of: International journal of simulation and process modelling. Volume 18:Number 4(2022)
• Page Start: 253
• Page End: 266
• Publication Date: 2023-01-11
• Subjects: road traffic accidents -- data mining -- analysis of the risk of accidents -- weighted naive Bayes -- improved k-means; Management -- Computer simulation -- Periodicals; Mathematical models -- Periodicals; Operations research -- Periodicals; Simulation methods -- Periodicals; 003.05
• Languages: English
• ISSNs: 1740-2123

Abstract

According to the characteristics of road traffic accident data, two improved data mining methods are used to analyse the risk of accidents: nine accident-related factors are selected for discrete classification by weighted naive Bayes, the influence between factors is measured by weights and PMI thresholds, and the type of accident was predicted for a combination of factors. The accuracy of prediction increased from 83.98% to 87.02%. The traditional k-means algorithm is improved from three aspects: initial clustering centre, outlier point and distance measurement. Through these improvements, the computational complexity of clustering process is reduced and the clustering accuracy of accident-related factors is improved. On the one hand, the two methods can quantify the risk of accidents and facilitate the formulation of preventive measures; on the other hand, they can be used to improve the rationality of traffic safety evaluation.