Research Of Road Traffic Accident Prediction Based On Exponential Smooting And Markov Chain

With the pace of Reform and Opening,China's economy has achieved remarkable results,and the transportation industry has also experienced rapid development.The mileage of roads and the number of motor vehicles have increased year by year,but the road traffic accidents that followed have been alarming.Since the 21 st century,our country has paid more attention to traffic safety and adopted a series of policies and measures,resulting in significant traffic safety results.However,the current traffic safety situation in China is not optimistic,and it still has a long way to go.In this paper,exponential smoothing method is taken as the main line of research.By combining Markov chain and Dempster-shafer theory of evidence,three regional traffic accident prediction models are established.Using the same set of historical data,the three established models were used to make predictions,and the prediction accuracy of the model was evaluated.Finally,the model with the highest prediction accuracy is used to forecast the number of road traffic accident deaths in China from 2016 to 2020.The forecast results provide a theoretical basis for formulating traffic safety policies.First,this paper studies the theoretical basis of road traffic accident prediction.The basic principle of road traffic accident prediction is expounded,and how to choose a suitable road traffic accident prediction method and evaluation index is explored and analyzed.The applicability of exponential smoothing method and Markov chain in road traffic accident prediction is analyzed.Secondly,a road traffic accident prediction model based on exponential smoothing method was established.Based on the analysis of traffic accident statistical indicators,a suitable exponential smoothing forecasting model was established,including model classification,smooth initial value determination,and smooth coefficient selection.Among them,the selection method of smoothing coefficient is mainly studied,and a new selection method of smoothing coefficient,Levenberg-Marquardt algorithm,is proposed,and the practicability of the algorithm is verified.Then,on the basis of exponential smoothing prediction,combined with Markov chain,a road traffic accident combination model based on exponential smoothing method and Markov chain is established,which covers the basic idea of the combined model,the construction of the model,and demonstrates through examples.The use of the model and the predictive effect.Finally,the defects of the classical Markov chain are analyzed,and the DS evidence theory is introduced to solve the defect.An improved combination model is established.Taking the number of road traffic accident deaths in China from 1991 to 2012 as the data source,the three prediction models established in this paper were calibrated,and the calibrated models were used to predict the number of deaths in road traffic accidents in China during 2013-2015,and the results were analyzed and compared.Selecting the model with the highest prediction accuracy predicts the development trend of China's future road traffic accidents,providing a theoretical basis for the formulation of traffic safety policies.