Macro Situation Analysis And Prediction Of Road Traffic Accidents Based On Seasonal Time Series

The macroscopic situation prediction of road traffic accidents is a relatively new research direction compared with the analysis of traditional road traffic accident micro-causes.Through a global prediction and analysis of the total amount of road traffic accidents,it is found that the macroscopic laws can be used for traffic Provide a reference for the safety prevention strategy and management objectives.The main research content of this article includes the following three aspects:?1?Aiming at the problem that the traditional time series model cannot solve all the uncertain factors that cause the prediction interval to be too narrow and the prediction accuracy is not high,this paper refers to the improved ETS model?Bagged ETS?based on the Bootstrap algorithm and the Bagging algorithm.The number of road traffic accidents counted in 24 hours is the main body of the study,which is divided into a training set and a test set.An exponential smoothing model,an ARIMA model,and an improved ETS model based on the training set are established and predicted,and the prediction results are compared with the test set.The study found that the Bagged ETS model has the highest prediction accuracy in the monthly accident number prediction.Among the 24h accident number prediction,ARIMA?1,1,0??2,1,0?₂₄has the highest prediction accuracy,and the prediction accuracy of the Bagged ETS model is higher than that of the ETS model.Therefore,the improvement of the ETS model is effective,but its applicability has limitations.?2?Based on the best prediction model,the number of road traffic accidents in China from 2014 to 2016 is divided into months and 24 hours respectively,and the following rules are obtained:the number of road traffic accidents divided into months and 24 hours has obvious seasonality and trend.It is empirically obtained that the number of road traffic accidents in China has decreased year by year and entered a period of slack,indicating that the effect of traffic safety measures implemented in China in the past ten years is better and can provide a theoretical basis for formulating traffic safety management goals.Through time series analysis and prediction,it is found that the road traffic accidents show certain changes in monthly statistics and 24h statistics:comprehensively,the high incidence months of road traffic accidents are April,August,September,November and In December,the low-incidence months are February and March,of which the highest month of accidents is December and the lowest months are February and March;the high-risk periods of road traffic accidents are 7:00,11:00,14:00,17:00 and At 18:00,the low-incidence period is from 23:00 to3:00,with the highest point of accident occurring at 18:00 and the lowest point of time at 3:00.These potential laws can provide a basis for transportation related departments to propose traffic prevention measures and formulate laws and regulations.?3?The analysis shows that the accident rate of 10,000 vehicles in 31 provinces in China from 2003 to 2013 has an obvious trend.Therefore,the damping trend model in the exponential smoothing method is introduced.By adjusting the damping parameters,the predicted value of each province is positive.Predict the accident rate of 10,000 cars in the typical ten provinces in the next five years,and draw the following rules:During the next five years?2014?2018?,the accident rate of 10,000 cars in Anhui,Hainan,and Guangdong provinces is higher,of which Anhui Province The rate of decline is relatively fast,Hainan and Guangdong are basically stable,and the rate of decline is very low;the accident rate of 10,000 cars in Shanghai,Hebei,and Beijing is low,and they are all declining at a relatively low rate,and Shanghai has fallen to the lowest.The above laws can provide references for provinces to formulate traffic policies and traffic safety management objectives.