、去均值 2、计算协方差矩阵及其特征值和特征向量 3、计算协方差矩阵的特征值大于阈值的个数 4、降序排列特征值 5、去掉较小的特征值 6、去掉较大的特征值(一般没有这一步） 7、合并选择的特征值 8、选择相应的特征值和特征向量 9、计算白化矩阵 10、提取主分量 -To mean the number of calculated co-covariance matrix and its eigenvalues ​ ​ and eigenvectors to calculate the covariance matrix characterized by a value greater than the threshold value 4, descending characteristic value 5, remove the smaller eigenvalue 6, to to swap large characteristic value (usually not this step) 7, merge the selected feature value of 8, to select the corresponding eigenvalues ​ ​ and eigenvectors 9, calculating the primary component of the whitening matrix 10, extraction