youtor.org
Introduction to Principal Component Analysis
#############################
Video Source: www.youtube.com/watch?v=_bSMW1Q9_Ks
An introduction to principal component analysis and covariance matrices. • Pre-requisites: • Matrix operations • References: • In my video I compare PCA to a sequence of linear regressions, to give a rough intuitive understanding of what PCA looks like. However, before I get objections from the purists about this comparison, I should mention that there is a difference in how linear regression minimizes error compared to how PCA does it. In summary, linear regression minimizes error perpendicularly to the independent variable's axis, while PCA minimizes error perpendicularly to the vector itself. For further explanation, read the answer to the following stackexchange question: • http://stats.stackexchange.com/questi... • Details about eigendecomposition algorithms: • Jacobi method: • http://www.fizyka.umk.pl/~jacek/docs/... • http://homepages.dcc.ufmg.br/~assunca... • http://math.fullerton.edu/mathews/n20... • QR Decomposition method: • http://www.inf.ethz.ch/personal/gande... • http://www.stat.wisc.edu/~larget/math...
#############################
