Research On Induced Ordered Weighted Geometric Average Combination Forecasting Model Of Highway Passenger Volume

Passenger traffic volume forecast is a necessary part of road network planning andbasis of calculating highway cost-effective. Usually, regression analysis method,exponential smoothing method, gray model, the neural network and the combinationprediction method are widely used to forecast highway passenger traffic volume. It isconsiderably difficult and complex to accurately describe changes and development lawof highway passenger volume for the volume is affected by various multi-level factors.Application of only one traditional method can just provide a limited accuracy ofpredictions. Bates and Granger first proposed combination of prediction methods in the1960s, and then many scholars compared combination forecast method with theprevious single prediction method to come to a conclusion that the combinationforecasting model greatly promoted prediction accuracy. Now research on combinationforecasting method has developed into an important study direction in prediction field.In order to improve prediction accuracy of highway passenger transportationvolume, the paper firstly considers various factors which have an impact on highwaypassenger traffic to classify these factors into different kinds, and then use grayrelational analysis method to rank these influencing factors associated with highwaypassenger traffic from largest to smallest. The paper selects the fit model to express andforecast the change of influencing factors. Secondly, the paper uses IOWGA operator tocombine three exponential smoothing model, GM (1,1) forecast model, multivariateregression model and BP neural network to establish a combination forecasting modelon the basis of existing passenger volume forecasting models. Finally, the paper useshighway passenger traffic and the influencing factors of historical data of HeilongjiangProvince as the respective example to establish a single prediction model andcombination prediction model of highway passenger volume. According to thecompared calculation, it can be obtained that the combination model gets the smallererror and better prediction accuracy, so that the model can be used as an effectivemethod of forecasting highway passenger volume. Then the paper analyzes future trendsof Heilongjiang Province highway passenger traffic to provide reference values fortransport planning.