A prediction Model of the Number of Taxicabs Based on Wavelet Neural Network

Wavelet neural network is superior to traditional neural networks in the fact that the wavelet functions have good time-frequency localization characteristics and fast decay. To improve the prediction accuracy of the number of taxicabs, in this paper, wavelet neural network was used to approximate the nonlinearity between taxi number and its influence factors. And we also apply wavelet neural network to model each influence factor to obtain their future values. Then inputting the predictive values of the influence factors to the model of taxicab number, the number of taxicabs in the future will be achieved. Simulations show that the wavelet neural network model of the number of taxicabs has higher prediction accuracy than the BP neural network model.