On Jan 22 2012, Egon Willighagen wrote: > On Sun, Jan 22, 2012 at 4:37 PM, Peter Murray-Rust <pm...@cam.ac.uk> wrote: > > On Sun, Jan 22, 2012 at 2:25 PM, Karol M. Langner <karol.lang...@gmail.com> > >> They sure are independent, if that's what you mean by orthogonal, > >> but many people would claim they are correlated (at least statistically). > >> > > Yes - they are independent variables which may or may not be statistically > > indepdent. In Factor Analysis (allied to principal components analysis) the > > algorithm tries to rotate orthogonal axes onto independent (non-orthogonal) > > axes. Then to select the independent non-orthogonal axes which are closest > > to orthogonality. I think we should reserve orthogonality for its precise > > meaning (right-angles in multidimensional parameter space) > > OK, OK, ... I meant to say: > > They span a 2D orthogonal space that describes in which things occupy > an area with large variance in both directions, where both access > point pretty much in the direction of the two variables we were > talking about. > > Egon
So in this terminology, the variance of prestige in OA is not large enough. -- written by Karol M. Langner Mon Jan 23 13:26:17 CET 2012 ------------------------------------------------------------------------------ Try before you buy = See our experts in action! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-dev2 _______________________________________________ Blueobelisk-discuss mailing list Blueobelisk-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/blueobelisk-discuss