The singular value decomposition (SVD) allows one to re-write a given matrix as a sum of rank one matrices. Specifically, using the SVD, one may re-write a given matrix A as follows:
,
where is the transpose of the i-th column of V. Further, the Eckart-Young-Mirsky theorem proves that the best rank k approximation to the matrix A is found by summing only the first k elements of the right-hand sum.
Read more