Research on Subway Passenger Flow Combination Prediction Model Based on RBF Neural Networks and LSSVM

摘要

In view of the subway passenger flow's random problem, nonlinear problem and so on, in order to predict the subway passenger volume more accurately, this paper designs a kind of parallel variable coefficient weighted combination prediction model based on Radial Basis Function Neural Network and Least Squares Support Vector machines. In this method, firstly, the original data is preprocessed. Then, this paper respectively sets up RBF Neural Network and LSSVM prediction model for training and calculating the weighting coefficient with the results of the training. Finally this paper separately does two kinds of models' prediction, and weights to get results. This article uses 2012 passenger flow data of Beijing DONGZHIMEN Station for experiments, which shows that the result of combination prediction model is more accurate than the result of single prediction model.