Application Research Of Combined Model Based On GIOWA Operator In Railway Passenger Traffic Forecasting

With the rapid development of China's economy,the connection between cities is getting closer,China has entered a period of large-scale investment and construction of railway passenger transport lines,especially high-speed railways,which has effectively alleviated the current tense situation of passenger transport capacity.In terms of railway transportation,forecasting traffic volume is an important task that must be completed before the investment and construction of railway planning.The accurate prediction of railway passenger volume plays an important role in the economic evaluation of railway construction projects,the allocation of national resources,the adjustment of railway internal investment structure,and the management of operations and so on.Therefore,how to make scientific and reasonable prediction of railway passenger volume has become an important research topic in the transportation field.Through the analysis and summary of the research status of railway passenger volume forecast at home and abroad,it is considerably difficult and complex to accurately describe changes and development laws of railway passenger volume for the volume is affected by various multi-level factors,application of only one forecasting method can just provide a limited accuracy of predictions.The related research shows that the combined prediction method obtained by combining different single prediction methods using GIOWA operator can effectively improve the overall prediction accuracy of the model.The model offsets defect of the fixed weighting of the traditional combined prediction method by sequentially weighting the fitting precision of each single forecasting method at each time point in the sample interval,thus effectively improving the predictive performance of the model.Firstly,according to the advantages and disadvantages and applicability of each single prediction model and the selection principles of single model in combined prediction method,combined with the development changes of railway passenger volume,this paper selecting the ARIMA model,partial least squares regression model and GRNN network model as the single predictive model for forecasting and analyzing railway passenger volume.Secondly,classifies and analyzes a number of factors affecting railway passenger volume,and finally determines the factors related to railway passenger volume by using the correlation coefficient method.And then,introducing the generalized induced ordered weighted averaging operator GIOWA and selecting IOWA,IOWGA,IOWHA operators,establishes the combination forecast models in railway passenger volume forecasting.In the process of assembling the index information,the combined prediction model considers the importance degree of the index set itself and the accuracy of the single-model prediction result.By integrating the information of each single-prediction model,the prediction error of the single-item model is effectively reduced.Finally,based on the historical data of China's railway passenger traffic and various influencing factors,establishes time series-based GIOWA combined prediction model and GIOWA combined prediction models based on multiple influencing factors.And then,the average absolute error,the mean square error and the average absolute percentage error are selected as the evaluation indicators to judge the accuracy of each model's prediction results.According to the compared calculation,the errors of the combined prediction models are smaller than the single prediction models and traditional prediction models,and the prediction accuracy is higher.Among them,the GIOWA combination prediction model based on the influencing factors has higher prediction accuracy,which verifies the validity and practicability of the combined prediction model proposed in this paper,thus providing an effective reference and consultation for traffic planning.