Method of nonlinear character extraction in communication traffic based on KPCA
一种基于核主元分析的话务量特征提取方法*

In view of the communication traffic feature, this paper presented a method of nonlinear character extraction based on kernel function principal component analysis(KPCA). Nonlinear character extracted by KPCA reflectd the complex relationship between original input and output data and simplified the array dimension of input data. By comparing simulation results, the prediction model based on KPCA-RBFNN has better ability to deal with nonlinear data than that prediction model based on PCA-RBFNN. The experimental results show that this method is very effective.