核函数主成分分析，用于数据的特征提取，对于训练样本的降维有较好的效果-Kernel principal component analysis, feature extraction for data, which can effectively reduce the dimension of training samples, the better