RPCA Procedure

The RPCA procedure implements robust principal component analysis (RPCA) in SAS Viya . RPCA can be used in many areas, including image processing, latent semantic indexing, ranking, and matrix completion (Candès et al. 2011).

The RPCA procedure decomposes an input matrix into a sum of two matrices: a low-rank matrix and a sparse matrix. You can use the low-rank matrix to do feature extraction and use the sparse matrix to detect anomalies.

Robustness in RPCA comes from the property that the principal components are computed from observations after removing the outliers—that is, from the low-rank matrix. Many applications of RPCA focus on the sparse matrix. One example is the extraction of moving objects from the background in surveillance videos.

PROC RPCA stores the results of the input matrix decomposition in output tables that are produced by the OUTLOWRANK=, OUTSPARSE=, and OUTERROR= options. Also, PROC RPCA stores the results of the low-rank matrix decomposition in output tables that are produced by the OUTDECOMP statement.