Research On The Characteristics Of Urban Traffic Accident Drivers Based On Data Mining

With the rapid development of social economy,the rapid urbanization and increase of vehicle ownership,urban road traffic safety problems are becoming more and more prominent Compared with the developed countries,the casualties and property losses caused by road traffic accidents are higher in China.There are many factors that can lead to traffic accidents.Human factors are the main factors causing traffic accidents,and the driver is dominant in human factors.This paper introduced the data mining technology,and introduced the related concepts of statistical analysis,clustering analysis and association rules and other data mining methods.Based on the traffic accident data of one city in China in 2015,the characteristics of urban traffic accident drivers are studied.First,the drivers' age,driving age,gender characteristics and other non-driver characteristics were analyzed,and the factors were analyzed comprehensively.By analyzing the driver's age distribution,driving age distribution and gender distribution,it was found that:(1)the probability of accident of drivers between 25 and 31 years old was 25.83%.(2)junior-driving drivers are high incidence of traffic accidents.As many as 35.77% of people have been driving for less than five years.(3)male driver's number of accidents was 91.72%,which is far higher than female driver's.The traditional gray definite weighted clustering algorithm is improved.The improved Logistic definite weighted clustering algorithm is used to analyze the driver accident tendency and use MATLAB to achieve the algorithm.Clustering results show that male driver's aged between 25 and 31 who have been driving for less than five years and male drivers who have been driving for between 6 and 15 years are the most likely to have accidents.The classification utility is used to evaluate the accuracy of the two clustering algorithms,and the CU values of the two clustering algorithms are calculated respectively.The results show that the CU of Logistic definite weighted clustering algorithm is 5.05% higher than that of traditional gray definite weightedclustering algorithm,which proves the superiority and effectiveness of the improved algorithm.Finally,the association rule's Apriori algorithm is used to analyze the driver's characteristics and other dimension data in the traffic accident.The algorithm is implemented by MATLAB,and the potential and meaningful association rules are obtained,such as 60 to 70 years old age accident driver have rich colors of cars,and the age of the driver in the 16 pm to 22 pm accident rate is high.Can be multi-angle,multi-level and more comprehensive study of urban traffic accident driver characteristics based on data mining technology.And dig deep into the potential causes of traffic accidents,what can provide theoretical basis and decision support for traffic safety management departments.