weight matrix from parcial warps doubt

[morphmet]

-------- Original Message --------
Subject: weight matrix from parcial warps doubt
Date: Fri, 4 Sep 2009 12:19:45 -0700 (PDT)
From: mizi...@rc.unesp.br
To: morphmet@morphometrics.org

Hello everybody!.

My name is Michele, I’m an undergraduate student from Brazil and I’m
working with geometric morphometrics of an endemic group of anurans from
Brazilian Atlantic Forest.

I made my analysis on tpsRelw, but I have a doubt about the weight 
matrix that I obtain from the parcial warps analysis. I didn’t 
understand how can I interpret  the axes generated by this matrix. I 
have 10 landmarks, so the matrix has 10x2-4=16 axis that represents the 
variation on shape. But, how can I associate this axis with my 
landmarks? Which landmark corresponds to each axis of variation? Is 
there a sheet where I can find these values? In the report  I didn’t 
find the value of each landmark for each axis (as a common PCA usually 
generates).. I only found one value for each landmark, but I don’t know 
 which axis I’ts referred to.

When I made my PCA with this matrix using another program (Statistica), 
I had the same problem. I couldn’t understand which landmark was related 
to the axis of variation of the PCA, because the values that I’ve 
obtained for each axis of the PCA was representing the importance of the 
axis of the weigth matrix (that, by the way, represents the axis of 
shape variation, not the landmarks).

I’ve read in many papers that I can use the weight matrix to do PCA in
another programs, but in any of them I could understand how to make the
correlation between the axis of shape variation and the landmarks.

Could you, please, help me to understand it?

Thanks a lot

Michele Fernandes Gonçalves
Undergraduate student

Universidade Estadual Paulista Júlio de Mesquita Filho – Unesp
São Paulo, SP, Brazil




--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org