Multi Cross-section Short-time Forecast Of High-speed Railway Passenger Flow Based On Ensemble Empirical Mode Decomposition-Multi Cross-section Gray Support Vector Machine

In the 20th century, high-speed railway in China has made remarkable achievements. China has been the country with the longest operating mileage and the biggest under-construction scale of high-speed railway in the world. Along with the rapid development of high-speed railway, the concept that "market demand decides transport supply" will become the norm for high-speed railway operation management. The short-term forecasting results of high-speed railway passenger flow are not only the basis of making train plans, but also the foundation of the ticket allocation and the earnings management. The short-term forecasting of high-speed railway passenger flow will be more and more important in the future. The existing passenger flow prediction methods still have improved space on the prediction precision. Therefore, it has the important practical significance to study the short-term forecasting methods of high-speed railway passenger flow.Based on the principle that passenger flow of multi cross-section has correlation, this paper explores and studies the problem of multi cross-section short-time forecast of high speed railway passenger flow. Firstly, on the basis of reviewing the related research about short-term passenger flow forecasting at home and abroad, we build the research framework and define the related concepts. Secondly, the related characteristics of high-speed railway passenger flow have been analyzed in depth in this paper, and summarized the short-term forecasting models and methods from the parameters, the parameters and combination forecasting three aspects, lay theoretical foundation for multi-section short-term traffic forecasting model. Thirdly, we applied the multidimensional time series analysis method to analyze the high speed rail passenger flow data, and judge the section passenger flow correlation on the base of the analysis results. Then the multi cross-section short-term forecasting method of passenger flow has been proposed and the multi cross-section short-term passenger flow forecasting model has also been constructed on the basic of EEMD-MGSVM. Finally, using the Wuhan-Guangzhou high-speed railway passenger flow data as the example to verify the multi cross-section short-term passenger flow forecasting model, and compared the forecasting results with forecasting results of single-section short-term passenger flow model. The results showed that multi cross-section short-term passenger flow forecasting model based on EEMD-MGSVM has better precision. The average percentage error (MAPE) is reduced by 5.69%, the average absolute deviation (MAD) is decreased by 10.76, the average mean square error (MSE) is decreased by 0.0117, and the square correlation is increased by 0.064. These results of the error evaluation verify the correctness and validity of the multi cross-section short-term passenger flow forecasting model constructed in this paper.