2 ms·
The furthest I have come was applying an SVD on a 2D Matrix to find the new axes of a skewed ellipsis, but I think I could follow for most of the part. Thanks.
by jansan 4y ago
The furthest I have come was applying an SVD on a 2D Matrix to find the new axes of a skewed ellipsis, but I think I could follow for most of the part. Thanks.
- epistasis 4y agoYeah, that's enough to visualize the idea. If you have a mx2 matrix, and you take each row as a point in 2D space, the compression means projecting all the points along the long dimension of the ellipse. If all the points are along a line, then the matrix is perfectly compressible.