Fits a PLSR model with the SIMPLS algorithm, modified to allow a weighted analysis.
simpls(X, Y, ncomp, stripped = FALSE, ...)A list containing the following components is returned:
an array of regression coefficients for 1, ...,
ncomp components. The dimensions of coefficients are
c(nvar, npred, ncomp) with nvar the number
of X variables and npred the number of variables to be
predicted in Y.
a matrix of scores.
a matrix of loadings.
a matrix of Y-scores.
a matrix of Y-loadings.
the projection matrix used to convert X to scores.
a vector of means of the X variables.
a vector of means of the Y variables.
an array of fitted values. The dimensions of
fitted.values are c(nobj, npred, ncomp) with
nobj the number samples and npred the number of
Y variables.
an array of regression residuals. It has the same
dimensions as fitted.values.
a vector with the amount of X-variance explained by each number of components.
a vector with the amount of Y-variance explained by each number of components.
Total variance in X.
Total variance in Y.
If stripped is TRUE, only the components
coefficients, Xmeans and Ymeans, Xvar and
Yvar, and Xtotvar and Ytotvar are returned.
a matrix of observations. NAs and Infs are not
allowed.
a vector or matrix of responses. NAs and Infs
are not allowed.
the number of components to be used in the modelling.
logical. If TRUE the calculations are stripped
as much as possible for speed; this is meant for use with
cross-validation or simulations when only the coefficients are
needed. Defaults to FALSE.
other arguments. Currently ignored.
Based on simpls.fit by Ron Wehrens and
Bjorn-Helge Mevik, with simple modifications by Gavin L. Simpson.
This function is a modified version of
simpls.fit from package pls. Four
modification have been made:
The input matrices X and Y are not centered,
The scores (tt in the code) are not centered,
Added code to calculate the total variance in the Y
matrix, Ytotvar, and the variance in Y accounted for
by each PLS axis, Yvar (See Value below), and
Additional components are returned if argument stripped
is TRUE.
This function should not be called directly, but through
the generic function coca.
SIMPLS is much faster than the NIPALS algorithm, especially when the number of X variables increases, but gives slightly different results in the case of multivariate Y. SIMPLS truly maximises the covariance criterion. According to de Jong, the standard PLS2 algorithms lie closer to ordinary least-squares regression where a precise fit is sought; SIMPLS lies closer to PCR with stable predictions.
de Jong, S. (1993) SIMPLS: an alternative approach to partial least squares regression. Chemometrics and Intelligent Laboratory Systems, 18, 251--263.
coca