Welcome to my blog
I’m glad you are here. I plan to talk about …
Posts
-
Principal-Component-Analysis
First introductions to Principal component analysis are usually delivered via so-called maximum variance or minimum reprojection error formulations, both of with lead to eigenvalue problems. Another useful view is that of probabilistic PCA, which is the subject of this post. The probabilistic view offers a number of benefits and generalisations. Two useful sources of information are chapter 12.2 of Bishop’s book and chapter 10.7 of Mathematics for Machine Learning.
subscribe via RSS