A predictive analysis model research based on the Bayesian classifier: Taking United Kingdom Vehicle Accidents as an instance

As economic and social development has led to a high rate of growth in road traffic, the associated road safety issues arising from vehicle accidents have received increasing attention. Although data on vehicle accidents and the factors that may cause them are available for most countries or cities, there is a lack of mature papers that examine the relationship between the factors that may cause vehicle accidents and the severity levels of vehicle accidents so that targeted measures can be taken to investigate them. The article uses data on UK vehicle accidents from 2005–2015 and uses a plain Bayesian classifier to explore the association between factors that may influence accidents and the severity rating of the resulting vehicle accidents. The results from the Bayesian classifier show that the highest accident severity levels occur on weekdays for temporal factors, on one-way streets for spatial factors, and on roads with high daylight and 30km/h speed limits for environmental factors. According to the statistical results of the analysis, the traveling public and local governments can be reminded to formulate effective management measures such as prevention and staggered travel based on the research results. Our contribution: 1. A Bayesian classifier was built to analyze the relationship between accident factors and accident severity levels. 2. The study results can help local traffic planning authorities in the UK develop flexible prevention and management measures to reduce the frequency of vehicle accidents for the relevant factors in response to the combined road conditions of the day.