-------- Original Message -------- Subject: RE: Fourier Descriptors Date: Thu, 30 Oct 2008 11:53:03 -0700 (PDT) From: F. James Rohlf <[EMAIL PROTECTED]> Reply-To: <[EMAIL PROTECTED]> Organization: Stony Brook University To: <[email protected]> References: <[EMAIL PROTECTED]> The is no reason to use fewer harmonics as input to a PCA. One would like to capture shape variation at all spatial scales. For a MANOVA you will have the practical problem that the analysis is undefined if you have more variables than the pooled within-groups covariance matrix (total n minus the number of groups). ========================= F. James Rohlf Distinguished Professor, Stony Brook University http://life.bio.sunysb.edu/ee/rohlf
-----Original Message----- From: morphmet [mailto:[EMAIL PROTECTED] Sent: Thursday, October 30, 2008 7:02 AM To: morphmet Subject: Fourier Descriptors -------- Original Message -------- Subject: Fourier Descriptors Date: Wed, 29 Oct 2008 18:29:58 -0700 (PDT) From: Catherine Sinfield <[EMAIL PROTECTED]> Reply-To: [EMAIL PROTECTED] To: morphmetric organisation <[EMAIL PROTECTED]> Dear all, I have applied EFA to analyse shape using the software package SHAPE , I found the best resolution with the first 20 harmonics . This gives an output of 80 fourier descriptors (now normalised for size etc)- what criteria should I follow for selecting FDs, do I need to use all descriptors in analysis such as PCA or MANOVA? Thanks very much -- Replies will be sent to the list. For more information visit http://www.morphometrics.org
-- Replies will be sent to the list. For more information visit http://www.morphometrics.org
