ABSTRACT
Although rarely tackled at the undergraduate level, SVD is extremely useful, particularly in statistics and signal processing. In Lab 19, we looked at square matrices that are diagonalizable and orthogonally diagonalizable. An important fact about the diagonalization is the resulting diagonal matrix contains the eigenvalues of the original matrix on the main diagonal. However, not all matrices are diagonalizable. In this case you may look at singular value decomposition (SVD). If A is a m×n matrix, singular values, σj , are the square roots of the eigenvalues for the matrix ATA.
