Comparative Analysis On Prediction Methods Of Highway Passenger Transportation Volume

Highway passenger quantity is the important indicator reflecting the service level of transport enterprises for the national economy and people’s life, and is the important indicator for performance evaluation of the road transport organization. The highway passenger quantity can reflect the transportation result of transportation department and social passenger transport demand.This paper, under the background of current highway passenger transportation situation and historical data in Jilin Province and Changchun City, studies the relationship between passenger capacity and the related factors, forecast the passenger capacity by different models, has contrastive analysis, proving the reliable basis for road transport administrative department and transportation enterprise in Jilin Province and Changchun City to reasonably authorize, check operation plan and capacity, construct the road network scientifically, which has important practical significance to improve the management level and operation efficiency of urban public transportation. The concrete research content is as follows:(1) Select many factors relating to highway passenger capacity, according to the ACI criterion, divide the passenger capacity in our province and its related factors during 34 years form reform and opening-up into two stages:from 1978 to 1997 and from 1998 to 2011, establish Cobb-Dauglas production function, form which we can clearly understand the relationship between the passenger capacity and its related factors, and the development of passenger transportation career in Jilin Province.(2) Use the annual data of road passenger capacity of Changchun City in 2003-2012, establish the grey prediction model; the model, through a series of test, residual test, correlation test, posterior difference test, obtains the mean absolute percentage error MAPE=2.99%. The results show that the model belongs to the high precision prediction model, and is used to predict the highway passenger capacity of Changchun City in 2013 and 2014.(3) Use the quarterly data of road passenger capacity of Changchun City in 2003-2012; first of all, build the three-exponential smoothing model to have smooth predict. Secondly, due to the student flow, migrant workers flow, holiday, tourist flows, the passenger capacity changes periodically with the seasons, so this paper, according to the raw data of passenger traffic, calculates the seasonal index number of each season. Third, in the form of the product, use the seasonal index to amend the three-exponential smoothing model value, obtain a satisfactory prediction result with the absolute error between 1.42 and 36.34, and the mean absolute percentage error MAPE=3.8%; the results show that the model belongs to the high precision prediction model, and is used to predict the road passenger capacity in each quarter of Changchun City in 2013 and 2014.(4) Use the quarterly data of road passenger capacity of Changchun City in 2003-2012; first of all, establish the GARCH (1,1) model to have smooth prediction. Secondly, use the seasonal index to amend GARCH (1,1) model prediction value, get an amended predicted value and a residual. Third, use the residual to calculate the residual mean value, use the residual mean value to amend the amended predicted value to get the second amended predicted value, namely the final predicted value. The error analysis shows that the second amended residual is between-3.933 and 6.261, the average absolute percentage error is MAPE=0.69%; the prediction is the high precision prediction, and is used to predict the road passenger capacity in each quarter of Changchun City in 2013 and 2014.(5) Compares and analyzes the models built in-(3) and (4). The conclusion is that the model in (4) is superior to that in (3). And the passenger capacity is dropping year by year with the passage of time, which has to do with air, bullet train, high-speed rail, substation, and accords with the practical significance.