From owner-chemistry@ccl.net Sun Nov 22 13:32:00 2009 From: "Niels Johan Christensen njc/./life.ku.dk" To: CCL Subject: CCL: Svar: CCL: PLS implementation? Message-Id: <-40753-091122123708-4589-ZCu3eBnH8A98O2NDtLFodA _ server.ccl.net> X-Original-From: "Niels Johan Christensen" Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=US-ASCII Date: Sun, 22 Nov 2009 17:46:13 +0100 Mime-Version: 1.0 Sent to CCL by: "Niels Johan Christensen" [njc=-=life.ku.dk] Hi Joe, One option is to implement a PLS algorithm in your favorite programming = language. The NIPALS PLS algorithm is particularly simple to understand = and work with (~10 lines of code in Matlab).=20 http://www.chemometrics.se/editorial/june2002.html#nipals Best=20 Niels http://www.chemometrics.se/editorial/june2002.html#nipals >>> "Joe Leonard jleonard42-$-gmail.com" = 19-11-09 0:01 >>> Sent to CCL by: Joe Leonard [jleonard42|gmail.com] Folks, I would appreciate any pointers to PLS implementations that can =20 handle much larger number of X variables as compared to Y (to-be-fit) =20 variables. I recall that PLS was used in 3D QSAR applications because =20 of its stability in such cases, but it seems that individual =20 implementations differ in how this statement applies to them... 3D QSAR test sets w/ aligned compounds would also be of use - if =20 people are aware of such collections on-line. But the PLS engine(s) =20 are my primary search right now... Thanks in advance! Joe Leonard jleonard42]=3D[gmail.com -- I have fought a good fight, I have finished my course, I have kept the faith. EFL/11-6-09 -=3D This is automatically added to each message by the mailing script = =3D-http://www.ccl.net/cgi-bin/ccl/send_ccl_messageSubscribe/Unsubscribe:=20Job: http://www.ccl.net/jobs=20http://www.ccl.net/spammers.txt