Research And Implementation Of Traffic Conditions Prediction Based On Mobile Signaling Data

How to accurately,efficiently and cost-effectively predict road traffic conditions is an important issue in the process of construction and improvement of intelligent transportation systems.Compared with road traffic conditions data collected by traditional methods such as floating cars,magnetic coils and cameras,mobile phone signaling data has the advantages of low cost and wide coverage.It takes important research value and significance to use mobile phone signaling data to extract road traffic conditions and predict traffic conditions.Through the research on the generation principle of mobile phone signaling number and related technologies,in the process of acquiring road traffic conditions by using mobile phone signaling data,a cleaning method for conventional abnormal data and a cleaning algorithm for noise data are proposed.The obtained user travel trajectory data on the road adopts a speed threshold-based clustering algorithm to identify the mode of travel of the motor vehicle,thereby finally obtaining the speed characteristic characterizing the road condition.The advantages and disadvantages of the various road traffic condition prediction models and their applicable scenarios are analyzed in detail.The in-depth analysis and research on the basic concepts of road traffic conditions,the characteristics of road traffic conditions,the state of road traffic conditions and the connotation of road traffic conditions are analyzed.Based on this,a combined forecasting model is proposed and implemented.The combined prediction model consists of an autoregressive integrated moving average(ARIMA)model based on statistical theory and a support vector machine(SVM)model based on intelligent theory.In the specific forecasting process,different predictive models are selected by judging the characteristics of road traffic conditions to achieve road traffic conditions are predicted.The support vector machine prediction model,the autoregressive integrated moving average prediction model and the combined prediction model based on the two models are tested respectively,and the test results are compared and analyzed.The results show that the combined prediction model has better prediction accuracy than the single prediction model.The combined prediction models can better adapt to the linear and nonlinear characteristics of road traffic conditions,and perform better in accuracy and robustness.