The real-time prediction of passenger flow in subway station passage sections is of great significance for operators to obtain real-time data support, ensure the safety of passenger flow in the station, and realize intelligent passenger transportation management. This paper adopts a real-time passenger flow forecasting model based on wavelet variable weight combination. Firstly, the trend and volatility of passenger flow data is decomposed and reconstructed by wavelet analysis method; secondly, Use two basic prediction models of SVM and RBF to make predictions, and add wavelet packet analysis method.; thirdly, the forecast results after the wavelet analysis obtained in the previous step are combined with variable weights to compare the prediction effects before and after the variable weight combination. The results show that the wavelet variable weight combination model, using wavelet analysis, SVM model and RBF model not only improves the prediction accuracy, but also improves the prediction stability.

CICTP