Application Of A Combination Of Traffic Forecasting Based On The Exponential Smoothing And ARIMA

Road traffic flow forecasting is an important task of modern traffic management and planning.Therefore, to accurately predict the future trends of traffic flow is also very important,what requires the study of the traffic forecast methods to improve prediction accuracy,to ensure correctness and rationality of the transportation projects to be built, and reduce road planning and design process and the waste of resources to improve social benefits.Through extensive collection of relevant literature,and combining with the results of wavelet analysis and exponential smoothing research, prior analysis of traffic forecasting methods is on the papers. The development and practical application of the theory of domestic and international traffic forecasts is discussed on the papers,analysis of the advantages and disadvantages of gray prediction theory, genetic algorithms, neural networks, etc are discussed on the papers.On this basis, there is a certain randomness, nonlinear changes by analyzing the traffic characteristics of historical data, anomalies appear in the data signal acquisition, the identification and correction method is proposed based on anomalies in the data signal acquisition occurs, summarizes the different methods of identification and correction of abnormal data, and the practical application of these methods to correct abnormal traffic data sequence data.In addition, getting related data of Zhengshao highway road traffic were analyzed on the basis of the current traffic forecasting methods and forecasting results, to further explore the feasibility of an innovative approach.According to the calculation principles of exponential smoothing, combined with data collected transnatured highway, due to the volume of traffic affected by many factors, by the factors on different scales separated (traffic hands on many factors), so that more apparent regularity making it easier to predict.This project will study the wavelet method which is introduced to predict traffic,and the use of predictive capabilities and three exponential smoothing wavelet multiscale analysis is presented and the wavelet multiscale exponential smoothing complex prediction is presented.Wavelet multiscale analysis theory and function is combined with exponential smoothing method reasonably together to grasp the nonlinear variation of traffic from essentially predict the effect,While using ARIMA model series by Spss software to predict traffic.Paper application portfolio through further research and forecasting methods, the linear combination of the above two methods effectively combine well verified by examples of wavelet multiscale exponential smoothing and ARIMA forecasting model combined forecasting model to predict traffic volume,which is very good and accurately predicts the future traffic speed Zhengshao long-term trends, enriches and develops the traffic forecast method, the operation and management of expressways provides important academic value and engineering value.