Producing computationally efficient KPCA-based feature extraction for classification problems
Yong Dong Xu, Chuang Lin, Wenda Zhao · Electronics Letters · 2010
An improvement to kernel principal component analysis (KPCA) to produce computationally efficient KPCA-based feature extraction is proposed. This improvement is applicable to all cases no matter whether the samples in the feature space have zero mean or not. Experiments on several benchmark datasets show that the improvement performs well in classification problems.