Forecast Of Road Freight Transportation Volume Based On KF-RBFNN Fusion Algorithm

Transportation is closely related to the national economy,and the development of the cargo transportation industry also affects China’s GDP level.The road freight transportation industry provides basic services for China’s economic development and drives regional economic development.Scientifically and reasonably predicting the volume of goods transported in a certain area in the future is crucial to transportation planning,construction,and management.At present,most of the forecast of cargo transportation volume is based on traditional forecasting methods.Since each forecasting method has its own forecasting mechanism and scope of use,the accuracy of the forecasting results is not high,so that the development trend of cargo transportation volume and the degree of impact on the economy cannot be accurately understood.Therefore,the use of computer tools to improve the accuracy of the forecast of freight traffic is an urgent problem to be solved.This paper proposes Kalman filtering and Gaussian kernel RBF neural network fusion algorithm(KF-RBFNN)to predict road freight traffic volume,based on the previous research results.The specific work is as follows:(1)This paper studies the existing domestic and foreign traffic volume prediction models,mainly including autoregressive moving average(ARIMA)prediction algorithm,gray prediction algorithm,neural network algorithm and fusion(or combination)prediction model,but most of them are based on traditional forecasting algorithm.By comparing and analyzing the existing prediction models,we can see that the prediction performance of the neural network model is superior to the traditional prediction model,and the fusion prediction method is superior to single prediction method.(2)This paper studies the prediction mechanism and derivation process of Kalman filter and RBF neural network,and studies the current situation of road freight transportation and freight transportation volume prediction.Based on the road freight transportation data,a fusion prediction model of Kalman filter and Gaussian kernel radial basis neural network was constructed,and the evaluation index of the prediction model was established.(3)This article selects China’s road freight transportation data from 2016 to 2019,and uses MATLAB for modeling and simulation.First,pre-process the original data,calculate the monthly data of freight volume,and normalize it.Then build Gaussian kernel RBF neural network model and KF-RBF fusion model,and substitute the pre-processed data for training and prediction.Finally,the evaluation index of the prediction model is used to judge the prediction performance of the model.Research shows that the prediction effect of KF-RBFNN is better.The fusion model uses Kalman filtering to compensate for the shortcomings of Gaussian kernel RBF neural network,making the model more robust,generalized,and the predicted value closer to the actual value Achieve the expected results,so as to provide reference for the transportation management department.