4 ms·
Singular value decomposition. It continually amazes me how many problems -- from noise reduction to Netflix prediction -- SVD can be usefully applied to. http
by gamache 16y ago
Singular value decomposition. It continually amazes me how many problems -- from noise reduction to Netflix prediction -- SVD can be usefully applied to.
http://en.wikipedia.org/wiki/Singular_value_decomposition http://en.wikipedia.org/wiki/Singular_value_decomposition
- cdavid 16y agoThe core ideas behind SVD and eigen values are very rich and profound. It has always fascinated me when I studying maths in undergrad. Someone else already mentioned compressed sensing, which expands on some of those ideas. Terrence Tao had a pretty good presentation on the topic: http://terrytao.files.wordpress.com/2009/08/compressed-sensing1.pdf http://terrytao.files.wordpress.com/2009/08/compressed-sensi...
- rdtsc 16y agoI still do not completely understand it. I see the math, but I see it as symbols that I can memorize, rather than something intuitive (like when I see a moving cursor through a state tree during a binary search for instance).