Input: The sample set $D=\left\{x_1,x_2,\cdots ,x_m\right\}$ ;

The dimension of the low-dimensional subspace is denoted as $d^{\prime }$ .

Output: The projection matrix W.

1: Centering all samples: $x_i\leftarrow x_i-\frac{1}{m}{\displaystyle {\sum }_{i=1}^m{x}_i}$ .

2: Calculate the covariance matrix $X{X}^{\text{T}}$ .

3: Perform eigenvalue decomposition on the covariance matrix $X{X}^{\text{T}}$ .

4: Take the eigenvectors corresponding to the $d^{\prime }$ largest eigenvalues $W=w_1,w_2,\cdots ,w_{d^{\prime }}$ .