Alexander Alexandrov
Saturday, March 12, 2016

Visualization of the SVD of a two-dimensional, real shearing matrix M. The SVD decomposes M into three simple transformations.
Let's generate some random data 40x10. Then add some pattern. So most of the rows from 5 up to 10 column are shifted by some value.
Pattern injected into the random data is not evident.
First singular vectors have catched most of the variance and pattern has become evident.