----- Forwarded message from Stefan Schlager
<zarquo...@gmail.com> -----
Date: Fri, 26 Apr 2013 09:37:46
+0200
From: Stefan Schlager <zarquo...@gmail.com>
Reply-To: Stefan Schlager <zarquo...@gmail.com>
Subject: Re:
2B-PLS vs PLSR
To: morphmet@morphometrics.org
simply put: while a PLS (from pls package) tries to minimize prediction error by iteratively selecting those linear combinations of the predictors with the highest covariation/correlation with the response variables, a 2block pls tries to find those linear combinations explaining the covariation between two sets of variables best. While the general approach is similar, the algorithms differ: the pls (NIPALS) is an iterative approach, while a 2Block pls uses a SVD of a block of the covariance matrix (much faster). The iterative approach results in better prediction quality, while the 2Block approach is mathematically simpler, faster and symmetric - i.e. predictor and response are interchangeable. In your case, however, where you want to predict skull shape from climate, I would prefer the pls approach. But be careful: use cross-validated predictions to estimate prediction quality, because generally, you need to be very cautious concerning overfitting. I used the pls-package in my dissertation to predict soft-tissue shape from hard-tissue shape, if you need help or code snippets, feel free to send me an email.
Best regards
Stefan
On 25/04/13 23:45, morphmet_modera...@morphometrics.org wrote:
----- Forwarded message from Rodrigo Lima <rodrigo.l...@mail.mcgill.ca> -----
Date: Wed, 24 Apr 2013 09:57:46 -0400
From: Rodrigo Lima <rodrigo.l...@mail.mcgill.ca>
Reply-To: Rodrigo Lima <rodrigo.l...@mail.mcgill.ca>
Subject: 2B-PLS vs PLSR
To: "morphmet@morphometrics.org" <morphmet@morphometrics.org>Dear morphometricians,
I have a question about the PLS analysis and I would be very thankful for any insight provided.
What is the difference between two-block PLS (as in Rohf and Corti 2000) and PLS regression (PLSR) implemented in the pls package in R (Mevik and Wehrens 2007)? I'm trying to relate skull shape to climatic variables, which one would be more appropriate in this case?
Thank you,
Rodrigo
----- End forwarded message -----
----- End forwarded message
-----
signature.asc
Description: PGP signature
