Analysis And Identification Of Spatial-Temporal Hotspots And Exploration Of Risk Factors For Traffic Accidents

Traffic accidents seriously affect people's lives and cause huge economic losses.It is of great practical significance to explore the hotspots and risk factors of traffic accidents.Based on the existing research at home and abroad,this study deeply analyses the spatial-temporal hotspots and risk factors of traffic accidents.Therefore,this study takes the traffic accident data of H city from 2013 to 2015 as the research object,and combines the traffic network data of H city to analyse the spatial and temporal distribution characteristics and risk factors of traffic accidents.Firstly,this study analyses the spatial and temporal attributes,and factor attributes of traffic accidents through the mathematical statistics method,so as to understand the distribution characteristics of traffic accidents in H city as a whole.Then,this study constructs a spatial-temporal network kernel density estimation model which takes into account the severity of traffic accidents,and analyses the spatial-temporal hotspots of traffic accident data.Furthermore,this study uses the Getis-Ord Gi*hotspot analysis method for spatial statistics to accurately identify the scope and boundary of the spatial-temporal hotspots of traffic accidents in H city.Finally,this study explores the structure learning method,parameter learning method and network reasoning method of Bayesian network.In the actual situation of complex traffic system,a large-scale traffic accident Bayesian network model is constructed to infer the probability of serious traffic accidents under different factors,and corresponding remedial suggestions are put forward.The results of this study are as follows:(1)From a spatial point of view,there is a high probability of a traffic accident in the administrative center areas in H City,and the spatial distribution shows a pattern of spreading along the southwest to northeast direction around the main trunk roads.From a temporal point of view,there is a high probability of traffic accidents during the early peak and evening peak in a day.And there is also a high probability of traffic accidents durning November and December of a year.In terms of weather factors,traffic accidents occur most in rainy days,secondly in snowy days,and least in sunny days.In terms of road grades,there is the biggest average loss of traffic accidents on the provincial roads,secondly on the national roads,and the smallest on the township roads.(2)The hotspots of traffic accidents show spatial distribution of "two belts and one center" in H City,i.e.the belt areas formed by the Xiaxu line and 101 provincial highway,and the administrative center area where the municipal government is located.The hotspots of traffic accidents show temporal distribution of regional feature correlation,i.e.the hotspots in industrial office areas occur in the morning peak period,the hotspots in residential areas occur in the evening peak period,and the hotspots in commercial consumption areas occur in the night period.The results of analysis and identification can accurately determine the spatial-temporal hotspots with serious traffic accidents,which will greatly save the manpower and material resources of the governance work.(3)Based on the Bayesian network model of traffic accidents in H city,it is concluded that there are some factors weigh havily for increasing the probability of serious traffic accidents.Firstly,the traffic accidents occur on sandstone and soil roads under snowy and rainy weather conditions.Then,the traffic accidents occur when hitting the fixtures on county roads,provincial roads,village roads.Finally,the traffic accidents occur in the early morning to dawn hours,involving motor vehicles and pedestrians on long-distance or mid-distance roads.Therefore,appropriate measures should be taken for early prevention and special treatment to reduce the probability of serious traffic accidents under these factors.In this study,analysis and identification of spatial-temporal hotspots as well as exploration of risk factors for traffic accidents were explored.The results accurately identify the scope and boundary of spatial-temporal hotspots of traffic accidents,and quantitatively infer the combination of risk factors affecting the severity of traffic accidents in the form of probability,which provide a comprehensive analysis perspective for traffic accidents.