Method, device and system for traffic flow prediction based on Bat algorithm

The embodiment of the invention discloses a method, device and system for traffic flow prediction based on the Bat algorithm. The method comprises: obtaining traffic flow data; employing a wavelets neural network traffic flow prediction model established in advance to perform processing and obtain a traffic flow prediction result, wherein the wavelets neural network traffic flow prediction model is formed through training based on the Bat algorithm, and the training process comprises calculation of initialization wavelets neural network parameters according to the historical data and the Bat algorithm; and employing the wavelets neural network and the historical data to perform training of the initialization wavelets neural network parameters to obtain the wavelets neural network traffic flow prediction model. Therefore, according to the embodiment of the invention, when the wavelets neural network traffic flow prediction model trained through adoption of the initialization wavelets neural network parameters