The Analysis And Prediction On Data Of Road Traffic Accidents In Sichuan Province

Road traffic safety is related to the safety of people traveling by car,the frequent occurrence of traffic accidents threatening the social stability of our country and the safety of people undefined property.The World Health Organization forecast that road traffic injuries will increase year by year.According to the data issued by the National Bureau of Statistics,the situation of traffic accidents in China is also very serious,and the serious traffic accident problems have a great influence on the economic development and social stability of our country.In order to alleviate and reduce the occurrence of traffic accidents and understand the factors causing road traffic accidents comprehensively and accurately,it is necessary to analyze and predict the historical road traffic data.In the early stage,taking the road traffic accidents in Sichuan Province as an example,the original road traffic accident data were collected,and the results caused by the road traffic accident were counted.Through the analysis of these data,the causes and factors of the accident were obtained.The factors strongly related to the occurrence of road traffic accidents are found,and the implicit laws are excavated.The algorithm based on double weight particle swarm optimization support vector regression is used to construct the traffic accident number prediction model to predict the number of accidents.Finally,the transplantation algorithm is used to develop the road traffic accident data analysis and prediction system,which provides convenience for the follow-up research.The main tasks as follows:(1)In view of the problem of data acquisition,the raw data of traffic accidents in Sichuan Province are collected,the results caused by road traffic accidents are counted,the data are analyzed briefly,and the data composition is analyzed from four angles of human,car,road and environment,respectively.It shows that the analysis of road traffic accidents needs multiple angles.(2)In view of the fact that the Apriori algorithm of association rules cannot get the required information and produce a large number of invalid and redundant data rules.The grey association analysis method is combined with Apriori algorithm,using grey relational analysis method to grasp the direction of interest can reduce the generation of redundant rules and facilitate the discovery of data laws.The method is applied to the road traffic accident data of Sichuan Province.Compared with the Apriori algorithm only,the experimental results show that the time of the algorithm is shortened by nearly 40% and the redundant rules are reduced by nearly 50%.The interesting relationship between road traffic accident data is excavated,and the effectiveness of the combination of grey correlation analysis and Apriori algorithm is proved.(3)In order to overcome the lack of data samples,unilaterally analyze traffic accidents,and predict the irrationality of units by year,an improved dual-weight particle swarm optimization algorithm is used to optimize the selection of support vector regression algorithm parameters,and to eliminate empirical acquisition.The blindness and randomness of the parameters enable the acquisition of optimal parameter combinations and improve the prediction accuracy of the model.Through the training and prediction of historical data of road traffic accidents in Chengdu from 2012 to 2016,the results are compared with the results of standard particle swarm optimization algorithm and mesh segmentation algorithm to optimize SVR parameters.The RMSE of the model is 0.408249 and MAE is 0.166667.The MAPE is 0.0277778 and the correlation coefficient is 98.5064%,which is better than the other two methods,which proves that it has better prediction ability.(4)A road traffic accident analysis and prediction subsystem was developed,which is part of the traffic accident analysis and prediction system.By transplanting algorithms such as Apriori and DWPSO-SVR,the corresponding algorithm can be executed by calling the corresponding method during coding.Finally,the functions of reading data,analyzing data,and predicting data are realized,and the data information is visually displayed in the form of a line chart and a histogram.The system allows government transport authorities to familiarize themselves with the factors that cause traffic accidents and the number of traffic accidents that may occur next month,and provides data support and decision-making guidance to policy makers.The data of road traffic accidents are statistically analyzed,and it is found that the occurrence of traffic accidents is the result of many aspects.By using the data of grey correlation analysis and Apriori algorithm,some factors strongly related to the occurrence of traffic accidents are found.And excavate their internal relationship.On this basis,the improved DWPSO-SVR algorithm is used to predict the data,and the evaluation index of the prediction results is better than the other two methods,which proves that the prediction effect is better.