Numerical Analysis 12.4.1 SVD and Compression

A simple application of the singular value decomposition is to compress the data in a matrix. By using a fraction of the storage space, you can often get a reasonable approximation to the original matrix. This could further be refined by the use of quantization.