Research On Reason Analysis And Prediction Of Road Traffic Accidents

In twenty-first Century,with the rapid development of economy,vehicle ownership and highway mileage have been increasing rapidly.This has led to great challenges to road traffic safety.Road traffic accidents have seriously threatened people’s lives and property safety.A lot of potential rules are hidden through the analysis of road traffic accident data,and it is of great significance to make corresponding measures to improve traffic safety according to these rules.Based on the data set of road traffic accidents in Leeds City,this paper studies the problem of road traffic accident analysis and prediction.The main work and research results are mainly reflected in the following aspects:(1)After data preprocessing to the original data of road traffic accidents,the attribute characteristics of the data set are analyzed by statistical analysis,and then the association rules are excavated based on the Apriori algorithm,and the rules of the mining are analyzed,and the condition attributes are analyzed from many aspects,such as people,cars,roads and environment.The impact of road traffic accidents.(2)The number of traffic accidents in consecutive days could be regarded as time series data.The EEMD algorithm of adaptive waveform matching extension is used for the first time to analyze the road traffic accident sequence under different time scales,and the continuous frequency and cycle of the decomposition results are calculated by Hilbert spectrum analysis,and the causes of the fluctuation are defined,the time series characteristics of the superposition components of the road intersection accidents are further obtained.From this,we can analyze the cause of the fluctuation of accident number and grasp the law of accident sequence at different time scales.(3)An road traffic accident prediction model based on the improved EEMD algorithm,which is combined with PSOSVR model and WNN model,is proposed,and the road traffic accident sequence under different time scales is predicted.After the decomposition of the road traffic accident sequence under each time scale,the prediction model based on all component input and the prediction model based on the combined component input are established according to the correlation and the power contribution rate of the original time series,and the prediction results of the components are obtained and the fusion prediction is made in the EEMD algorithm.The optimal model is selected through experiments to obtain accurate prediction results.