PCA for gradience of cost without zero mean shift

1 view
Skip to first unread message

Jiaji Zhou

unread,
Feb 21, 2013, 4:27:43 PM2/21/13
to physics-inspired-m...@googlegroups.com
Hi, Guys! 
I am not sure whether this is true. But I think if we do mean shift before PCA, just use the mean of all the gradience as the first principle component, and shift all the PCs by one. We will get the same thing of doing PCA without mean shift. 

Best
Jiaji (Robin) Zhou
Reply all
Reply to author
Forward
0 new messages