PCA is mathematically equivalent to eigenvalue decomposition of the data covariance matrix, or singular value decomposition (SVD) of the data matrix