The recent posts I've made regarding linear algebra have generated an unexpected amount of traffic. While I hadn't planned on writing as much as I did on the subject of the Singular Value Decomposition, what frustrates me is that I feel there's so much more that needs to be said. Especially the following:
1. The relationship between the SVD and the Moore-Penrose Pseudoinverse.
2. The relationship between SVD and PCA and dimensionality reduction.
3. The relationship between SVD and general least squares.
For now, I need to attend to other things and take a break. If you've been at all intrigued by the series so far, you may want to do more research on the three cases listed above.
Thanks a lot for your thoughts on the SVD so far! I would indeed appreciate a post on the relation between SVD and PCA.
ReplyDeleteThanks again!!
You're welcome. Hopefully, I can back to this soon. Mostly, I think I'm just waiting for some insight in how to best present it. I have a number of links to both PCA and SVD on my del.icio.us.
ReplyDelete