low-rank approximation

From Wikipedia, the free encyclopedia