Research Article

Wind Speed Forecasting by Wavelet Neural Networks: A Comparative Study

Table 1

Model parameters of the three wavelet neural networks and BP networks.

Test stationModelWavelet preprocessing moduleNeural network prediction module
Wavelet
function
Decomposition
coefficients
Structure of networkTransfer function of hidden layerTransfer function of output layerTraining algorithm

Test station 1 BPNN4-12-1*tansigpurelintrainlm
CWNN4-9-1Morletpurelintrainlm
LWNN , , , :  3-11-1tansigpurelintrainlm
db4 : 4-7-1tansigpurelintrainlm
: 4-12-1tansigpurelintrainlm
: 4-12-1tansigpurelintrainlm
LCWNN , , , : 3-9-1tansigpurelintrainlm
db4 : 4-7-1tansigpurelintrainlm
: 4-7-1tansigpurelintrainlm
: 4-12-1tansigpurelintrainlm

Test station 2 BPNN4-11-1tansigpurelintrainlm
CWNN3-8-1Morletpurelintrainlm
LWNN , , : 3-8-1tansigpurelintrainlm
db3 : 4-7-1tansigpurelintrainlm
: 4-12-1tansigpurelintrainlm
LCWNN , , : 3-9-1tansigpurelintrainlm
db3 : 3-10-1tansigpurelintrainlm
: 4-11-1tansigpurelintrainlm

* m-n-p denotes structure pattern of neural network. m, n, p are the numbers of the input, the hidden, and the output layer nodes, respectively.