Abstract：The existing grey models have the disadvantages of the higher fluctuation of data and distortion when using the first order linear differential whitening equation to fit traffic flow data. To improve the accuracy of short-term traffic flow prediction, a short-term traffic flow prediction model based on grey ELM neural network was proposed. Firstly, the short-term traffic flow data was accumulated by gray model to convert it into the long time traffic flow data and to reduce its randomness, which could effectively reduce the error caused by the fluctuation of the data. Then, ELM neural network was used to replace the first order linear differential whitening equation to predict the long-term traffic flow. Finally, the short-term traffic flow prediction results could be obtained by restoring the long-term traffic flow prediction results. And it improved the accuracy of prediction. The simulation results showed that the proposed method was an effective short-term traffic flow prediction method and was much more accurate than the existing methods.

Keyword：grey model;short-term traffic flow prediction;ELM neural network;first order linear whitening differential equation;