You are here: CTSPedia>CTSpedia Web>StatisticalTools>RScripts>PartialLeastSquaresR (14 Apr 2009, CTSpediaAdmin)EditAttach

MacroHeaderForm edit

Macro Name

Multivariate Partial Least Squares (R)

Description This function computes the K PLS components based on X & Y matrices and associated quantities. Y can be a vector or a matrix of multiple responses.
Macro Parameters Input argument(s)
X_s Matrix of predictors (genes)--standardized to mean 0 variance 1
variance 1 ("training X" of size n x p).
Y_s Matrix (or vector) of responses Y_s--also standardized
("training Y" of size nxL).
X_ps Matrix of predictor values (test/validation X) to construct
test components, X_ps (of size mxp).
lv The number of PLS components (lv).

Return(s):
T1 n x lv matrix of PLS training components based on training
information.
T2 m x lv matrix of PLS test components contructed from training
information.
PVEX 1 x lv vector of cummulative percent of X-variation explained.
PVEY 1 x lv vector of cummulative percent of Y-variation explained.
W p x lv matrix of X-weights.
B p x 1 vector of "regression coefficients".
V p x 1 vector of linear combination of sum of squares of X-weights
This is the so called 'VIP' (variable influence on projection)
(Wold 1994; SAS Institue Inc., 1999).

History:
Date/Last Modified: 10.28.01/07.19.01 (Danh V. Nguyen), Matlab code.
Translated to R in 2007 (Ying Chen).
Full testing of exact correspondence of results to Matlab, R and SAS
08.14.08 (D.V. Nguyen).
Programmer DanhNguyen?
Revision 1
Disclaimer History
Date/Last Modified: 10.28.01/07.19.01 (Danh V. Nguyen), Matlab code.
Translated to R in 2007 (Ying Chen).
Full testing of exact correspondence of results to Matlab, R and SAS
08.14.08 (D.V. Nguyen).

References
SAS Institute Inc. (1999). The PLS procedure. SAS/STAT User's Guide,
Version 8, Cary, NC, pp. 2693-2734.
Wold, S. (1994). PLS for multivariate linear modeling. In Waterbeemed,
H. (ed), Chemometric Methods in Molecular Design, Verlag-Chemie,
Germany, pp. 195-218.
Example(s)

Special Features

Called Macros This function calls other function(s): None.
Notes
  • X is standardized so rank(X)=n-1 so maximum value that lv can take is n-1.
    * Also, to get the correct test components the standardization of the test data X_p (denoted X_ps) needs to be done properly; i.e. based on column means and variance from the training data. X_ps is obtained by call to function STANDARDIZE_PRED.m.
See Also PartialLeastSquares
PartialLeastSquaresSAS

► Attachments

▼ Attachments

Topic attachments
I Attachment Action Size Date Who Comment
elser PLS.r manage 8.7 K 17 Mar 2009 - 14:42 CTSpediaAdmin PLS R Function
txttxt PLS_R.txt manage 8.7 K 17 Mar 2009 - 14:43 CTSpediaAdmin PLS R Function Text File
Edit | Attach | Print version | History: r3 < r2 < r1 | Backlinks | Raw View | Raw edit | More topic actions
Topic revision: r3 - 14 Apr 2009 - 13:39:57 - CTSpediaAdmin
 
Copyright &© by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding CTSPedia? Send feedback