version 1.2

DESCRIPTION

@@ -1,25 +1,22 @@
 Package: SCGLR
 Type: Package
 Title: Supervised Component Generalized Linear Regression (SCGLR)
-Version: 1.1
-Date: 2012-11-21
+Version: 1.2
+Date: 2014-02-21
 Author: Mortier F., Trottier C., Cornu G., Bry X.
-Maintainer: G. Cornu <gcornu@cirad.fr>
-Contact: F. mortier <fmortier@cirad.fr> and G. Cornu <gcornu@cirad.fr>
-        and C. Trottier <catherine.trottier@univ-montp3.fr>
+Maintainer: "G. Cornu" <gcornu@cirad.fr>
+Contact: "F. Mortier" <fmortier@cirad.fr> and "G. Cornu"
+        <gcornu@cirad.fr> and "C. Trottier"
+        <catherine.trottier@univ-montp3.fr> and "X. Bry"
+        <xavier.bry@univ-montp2.fr>
 Description: SCGLR extends the Fisher Scoring Algorithm so as to combine PLS
     regression with GLM estimation in the multivariate context.
-License: CeCILL-2
-Depends: Formula,expm,ROCR,ggplot2
-Imports: ggplot2,graphics,grid, stringr
-Suggests: parallel,testthat
+License: CeCILL-2 | GPL-2
+Depends: R (>= 3.0.0)
+Imports: Matrix,Formula,expm,graphics,ggplot2,grid,stringr,ROCR
+Suggests: parallel
 LazyData: yes
-Collate: 'data.r' 'f2x.r' 'scglr.r' 'utils.r' 'plot.r' 'kComponents.r'
-        'multivariateGlm.fit.r' 'oneComponent.r' 'critConvergence.r'
-        'multivariateFormula.r' 'scglrCrossVal.r'
-        'multivariatePredictGlm.r' 'infoCriterion.r' 'metric.r'
-        'multivariateGlm.r' 'print.r' 'summary.r'
-Packaged: 2013-08-01 10:54:11 UTC; cornu
+Packaged: 2014-03-18 09:42:05 UTC; cornu
 NeedsCompilation: no
 Repository: CRAN
-Date/Publication: 2013-08-01 13:58:25
+Date/Publication: 2014-03-18 11:44:18

MD5

@@ -1,52 +1,50 @@
-cd877f4d121fddca68ea27123994f83d *DESCRIPTION
-aa611912aad7c7c5b6ee97dbf9235114 *NAMESPACE
-d16247931f88ded16fb00afdb9e3afbf *R/critConvergence.r
-f5209b7d93d0cb2ff426ff79144019a4 *R/data.r
-832f9d9e53b88274bd970f66826d957e *R/f2x.r
-21ed50a3efc5a14b1064ac8049da47dc *R/infoCriterion.r
-35c862bdd3a5200df45e37e4b572ca25 *R/kComponents.r
-75a27bc0b0a4fff94aaca1ea37091b99 *R/metric.r
-d22fbf85bd9b227c39ef41c235983581 *R/multivariateFormula.r
-b9cb8f19ec765c9073892cc0449aebc0 *R/multivariateGlm.fit.r
-fc612f54373b2db4ca9626a8f55f984a *R/multivariateGlm.r
-4c82299d4fbd1e0ef6962cb9e73e4130 *R/multivariatePredictGlm.r
-e24a82814b7bc947d4f30c93efc1835f *R/oneComponent.r
-c863427f136e1c1118d15f0fd943fa55 *R/plot.r
-89d07a9b80905ec09ef9375e77354384 *R/print.r
-37331188923c048f0b74afbb142b1568 *R/scglr.r
-306c95f7503e9257328387518fccbc6f *R/scglrCrossVal.r
-f3ef6ed4b6e5f13f0ef09f6dec8d2ab6 *R/summary.r
-024c193656421f4ccd65eb81359c531f *R/utils.r
+22fafb0e29932530db904dc4ac769dd7 *DESCRIPTION
+5b276e12421ccba73aedfbcff1fec732 *NAMESPACE
+0e3e33861096acb17a23427da77d9780 *R/critConvergence.r
+4f9de2b0fb5d5f8c406eb0c4ea45679f *R/data.r
+e1545ea41fe2959052292b554d25d8ba *R/f2x.r
+416bd61514d52323209a62dbd8916160 *R/infoCriterion.r
+cdf86f32ce50c657a8999be3fbbe745a *R/kComponents.r
+58868746ee7485ae4a561700bfde1369 *R/metric.r
+632c86a07ea571b3d6fe36c841ec33f4 *R/multivariateFormula.r
+26ff07ca0c50f66a4007594d68b2f46b *R/multivariateGlm.fit.r
+3a5c41707e779dae512967e5583c3658 *R/multivariateGlm.r
+1623faa4d4209c3440489386c65f04fd *R/multivariatePredictGlm.r
+30464b02cb580afab80d14f59c3b7f02 *R/oneComponent.r
+23a23a4be562ec875a37286a1cf48605 *R/package.r
+70f8bba5c062a2f0a12832f1bd104277 *R/plot.r
+262c34c9eef94d2fb02bdfd82191a6cb *R/print.r
+79757bdb68af4becbcf1b1a8cabcaf5c *R/scglr.r
+ad21d512066a8c0f81e1493d2d2e6773 *R/scglrCrossVal.r
+b6a6a4fcd2d3a166d9c94c9ba06c6360 *R/summary.r
+bc6ce52c4c87329eb6e0aeb4dcdaf977 *R/utils.r
+dec380453004ab40ba23b0d57287da99 *build/vignette.rds
 b454e91b945abf3eaaa1d2ddf9c52cfc *data/genus.rda
 7228f976812aee4fdd60f7ac760b4a43 *demo/00Index
 7ff42f88c9d37c33feb1dea01efbeebe *demo/scglr.r
 725274da96c067732984971ed539f43d *demo/scglrCrossVal.r
 619f968ac2ad86861dcb639dcc703aa8 *demo/scglrPlots.r
-48f76694c3c4fb155d98846af530bf70 *demo/scglrPrediction.r
-e52b9d7a956182d4006ca7e141e67c0a *inst/doc/index.html
-d68073bb32e2ad6f6653784793d0eeee *inst/doc/scglrJss.pdf
+f42b1d5ded8aac352618c1d24ab5d607 *demo/scglrPrediction.r
+86646a041ca3f5e2be92496231802aae *inst/doc/scglrVignettes.R
+7439805d8b7c345345eb8b09adb229c0 *inst/doc/scglrVignettes.pdf
+16e8bebb00aab52f05f737a4d789c933 *inst/doc/scglrVignettes.rnw
 725274da96c067732984971ed539f43d *inst/examples/exCrossVal.r
 619f968ac2ad86861dcb639dcc703aa8 *inst/examples/exPlots.r
-48f76694c3c4fb155d98846af530bf70 *inst/examples/exPrediction.r
+f42b1d5ded8aac352618c1d24ab5d607 *inst/examples/exPrediction.r
 7ff42f88c9d37c33feb1dea01efbeebe *inst/examples/exScglr.r
-6ba2dcf16373185f6210a87b4412af9c *man/barplot.SCGLR.Rd
-0b2a3368dc311b11961d40383d5d9df4 *man/critConvergence.Rd
-817e3d77be9b91b141d178f0e7321593 *man/f2x.Rd
-e518eb01227aa85a575128573edf6209 *man/genus.Rd
-0e14990ab7b77cbaebda46965f521361 *man/infoCriterion.Rd
-06cdf85b5ea3d922cd4aa9384004d5cf *man/kComponents.Rd
-77d495773150608671851b92384577ab *man/metric.Rd
-01114b16f8f7d038602a7a6f7b3d4167 *man/metricBloc.Rd
-655b4573236a861a9759b57cd1e84ba3 *man/multivariateFormula.Rd
-19139893f117e69d8aa6d843ecc3260b *man/multivariateGlm.Rd
-3b62dc99b2ef035d1fc751d115bb32f0 *man/multivariateGlm.fit.Rd
-0cd40195714681edbf9791fa584eaf17 *man/multivariatePredictGlm.Rd
-d800e9139afd06158f43a9c3f2cdb8f9 *man/oneComponent.Rd
-2e5ecbec80d447b6f30e86d834b9c493 *man/pairs.SCGLR.Rd
-15b3fc8cc44677f6ea5161704bbe1dd4 *man/plot.SCGLR.Rd
-c9a14368af5383c4d175388cf3a1a71b *man/print.SCGLR.Rd
-9da8afeccf87037a24ef35aafc35f094 *man/scglr.Rd
-296356d50a40ccccf0cfa17fae762f4c *man/scglrCrossVal.Rd
-dcfdca164febe45cd241a4f7d4fd07e2 *man/summary.SCGLR.Rd
-ca2a3472f67fcc1315026280468579d6 *man/wtCenter.Rd
-2c2e17ac2e94d5ee2b6e6d1e47cc6d50 *man/wtScale.Rd
+efe4660e89d3f7a8a48942377006d5e7 *man/barplot.SCGLR.Rd
+d28e9af64ae418aae507b7f58bfc106b *man/critConvergence.Rd
+d541c67056c9b94c0c6e811c77f6d977 *man/genus.Rd
+84bf979373387ee46ecdae7dbbeade80 *man/infoCriterion.Rd
+6587703cc35632041588eec0a83c2782 *man/multivariateFormula.Rd
+3416d2207c1b015ca75148352cf0a6af *man/multivariateGlm.Rd
+c42a9863f30ecb8d0e149007158017db *man/multivariatePredictGlm.Rd
+8121d257d4e66c8a63a5194281c3f64d *man/pairs.SCGLR.Rd
+30250ccd180a7a7a27ea042ba70b0119 *man/plot.SCGLR.Rd
+3ecaa356e70e58defc59d8d7d274ad31 *man/print.SCGLR.Rd
+a3049d54125a480aea8ec0d4199e7fd7 *man/scglr-package.Rd
+11cfcd36bf73bcea83bf995e897bd2a9 *man/scglr.Rd
+87c6c1f825baf63b53c4f4983bf4ea68 *man/scglrCrossVal.Rd
+af77eb78cca2f9fdb5e7586307e1e883 *man/summary.SCGLR.Rd
+3088a2c384aac3c5287caca92e3cbf07 *vignettes/bibScglr.bib
+16e8bebb00aab52f05f737a4d789c933 *vignettes/scglrVignettes.rnw

NAMESPACE

@@ -2,23 +2,21 @@ S3method(barplot,SCGLR)
 S3method(pairs,SCGLR)
 S3method(plot,SCGLR)
 S3method(print,SCGLR)
+S3method(print,summary.SCGLR)
 S3method(summary,SCGLR)
 export(critConvergence)
-export(f2x)
 export(infoCriterion)
-export(kComponents)
-export(metric)
-export(metricBloc)
 export(multivariateFormula)
 export(multivariateGlm)
 export(multivariateGlm.fit)
 export(multivariatePredictGlm)
-export(oneComponent)
 export(scglr)
 export(scglrCrossVal)
-export(wtCenter)
-export(wtScale)
+import(Formula)
+import(ROCR)
+import(expm)
 import(ggplot2)
-import(graphics)
 import(grid)
 import(stringr)
+importFrom(Matrix,bdiag)
+importFrom(graphics,barplot)

R/critConvergence.r

@@ -1,14 +1,15 @@
-#' @title Auxiliary function tuning the convergence of scglr
-#' @description Function tuning the convergence of scglr
+#' @title Auxiliary for controlling SCGLR fitting
+#' @description Auxiliary function for \code{scglr} fitting used to 
+#' construct a convergence control argument.
 #' @export 
-#' @param tol positive convergence threshold
-#' @param maxit integer, maximum number of iterations
-#' @return a list containing elements named as the arguments
+#' @param tol positive convergence threshold.
+#' @param maxit integer, maximum number of iterations.
+#' @return a list containing elements named as the arguments.
 critConvergence <- function (tol = 1e-6, maxit = 50) 
 {
   if (!is.numeric(tol) || tol <= 0) 
     stop("value of 'tolerance' must be > 0")
   if (!is.numeric(maxit) || maxit <= 0) 
     stop("maximum number of iterations must be > 0")
   list(tol = tol, maxit = maxit)
-}
+}

R/data.r

@@ -1,25 +1,28 @@
-#' @title This is the data to be included into the package
+#' @title Sample dataset of abundance of genera in tropical moist forest
 #' @name genus
 #' @docType data
 #' @author CoForChange project 
-#' @references \url{http://www.coforchange.eu/fr/}
-#' @keywords data
+# @keywords datasets
 #' @note The use of this dataset for publication must make reference to the CoForChange project. 
-#' @description genus gives the abundance of 27 common tree genera in the tropical moist forest 
+#' @description Genus gives the abundance of 27 common tree genera in the tropical moist forest 
 #' of the Congo-Basin and 40 geo-referenced environmental variables on one thousand 8 by 8 km plots
 #'  (observations). Each plot's data was obtained by aggregating the data measured on a variable 
 #'  number of previously sampled 0.5 ha sub-plots. Geo-referenced environmental variables were 
 #'  used to describe the physical factors as well as vegetation characteristics. 
 #'  14 physical factors were used pertaining the description of topography, geology and rainfall 
-#'  of each plot. Vegetation is characterized through 16-days enhanced vegetation index (EVI) data. 
-#' @param gen1 to gen27 are the abundance of the 27 common genera on each plot
-#' @param altitude is the above-sea level in meters
-#' @param pluvio_yr is the mean annual rainfall
-#' @param forest forest type classified into seven classes
-#' @param pluvio_1 to pluvio_12 are the monthly rainfalls
-#' @param geology is a 5-level factor giving the geological substrate
-#' @param evi_1 to evi_23 are the means over 10 years of the enhanced vegetation indexes 
-#' for each successive 16 days period using  MODIS images
-#' @param lon and lat are the longitude/latitude of the plot centers
-#' @param surface is the surface of the sampled area contained in each 8 by 8 kms aggregated plot
+#'  of each plot. Vegetation is characterized through 16-days enhanced vegetation index (EVI) data.
+#' @source  \url{http://www.coforchange.eu/fr/}
+#' @format
+#' \tabular{ll}{
+#'    \code{gen1 to gen27} \tab abundance of the 27 common genera.\cr
+#'    \code{altitude} \tab above-sea level in meters.\cr
+#'    \code{pluvio_yr} \tab mean annual rainfall.\cr
+#'    \code{forest} \tab classified into seven classes.\cr
+#'    \code{pluvio_1 to pluvio_12} \tab monthly rainfalls.\cr
+#'    \code{geology} \tab 5-level geological substrate.\cr
+#'    \code{evi_1 to evi_23} \tab 16-days enhanced vegetation indexes.\cr
+#'    \code{lon and lat} \tab position of the plot centers.\cr
+#'    \code{surface} \tab sampled area.\cr
+#' }
+#' 
 NULL

R/f2x.r

@@ -1,14 +1,14 @@
-#' @title Regression coefficients conversion
-#' @description converts the regression coefficients associated with the components 
-#' into coefficients associated with the initial regressors
-#' @export 
-#' @param Xcr matrix of the standardized regressors
-#' @param centerx vector of the regressors' means
-#' @param invsqrtm metric matrix M^(-1/2)
-#' @param gamma (generalized) regression coefficients associated with the components 
-#' @param u matrix of the component coordinates on Xcr (component = Xcr * u) 
-#' @param comp matrix of the components
-#' @return beta matrix containing the regression coefficients associated with the regressors X
+# @title Regression coefficients conversion
+# @description converts the regression coefficients associated with the components 
+# into coefficients associated with the initial regressors
+# @export 
+# @param Xcr matrix of the standardized regressors
+# @param centerx vector of the regressors' means
+# @param invsqrtm metric matrix M^(-1/2)
+# @param gamma (generalized) regression coefficients associated with the components 
+# @param u matrix of the component coordinates on Xcr (component = Xcr * u) 
+# @param comp matrix of the components
+# @return beta matrix containing the regression coefficients associated with the regressors X
 f2x <- function(Xcr,centerx,invsqrtm,gamma,u,comp)
 {
   n <- dim(Xcr)[1]

R/infoCriterion.r

@@ -1,17 +1,17 @@
 #' @title Function that calculates cross-validation selection criteria
 #' @export
-#' @param ynew data matrix corresponding to the observations used as test sample
-#' @param pred predicted value of the linear predictor obtained from Xnew and the estimated parameters
+#' @param ynew data matrix corresponding to the observations used as test sample.
+#' @param pred predicted value of the linear predictor obtained from Xnew and the estimated parameters.
 #' @param family a vector of the same length as the number of responses containing characters 
 #' identifying the distribution families of the dependent variables.
 #' "bernoulli", "binomial", "poisson" or "gaussian" are allowed.
 #' @param type information criterion used. Likelihood, aic, bic, aicc or
-#'  Mean Square Prediction Error (mspe) are defined. auc also defined for Bernoulli cases only.
+#'  Mean Square Prediction Error (mspe) are defined. Area Under ROC Curve (auc) also defined for Bernoulli cases only.
 #' @param size describes the number of trials for the binomial dependent variables. 
 #' A (number of statistical units * number of binomial dependent variables) matrix is expected.
-#' @param npar number of parameters used for penalisation
+#' @param npar number of parameters used for penalisation.
 #' @return a matrix containing the criterion value for each dependent variable (row) 
-#' and each number of components (column)
+#' and each number of components (column).
 infoCriterion <- function(ynew,pred,family,type,size=NULL,npar=0)
 {
   if(type=="auc")

R/kComponents.r

@@ -1,25 +1,26 @@
-#' @title Estimation algorithm for K components
-#' @description calculates the K components by iteratively calling function oneComponent 
-#' @export
-#' @param X matrix (n*p) containing the standardized covariates
-#' @param Y matrix (n*q) containing dependent variables
-#' @param AX matrix of additional covariates used in the generalized regression but not entering the linear 
-#' combinations giving components
-#' @param K integer specifying the number of components
-#' @param family a vector of the same length as the number of responses containing characters 
-#' identifying the distribution families of the dependent variables.
-#' "bernoulli", "binomial", "poisson" or "gaussian" are allowed.
-#' @param size matrix of size statistical units * number of binomial responses, giving the number of trials 
-#' for binomial dependent variables. 
-#' @param offset used for the poisson dependent variables.
-#' A vector or a matrix of size: number of observations * number of Poisson dependent variables is expected
-#' @param crit a list of maxit and tol, default is 50 and 10e-6. If responses are bernoulli variables only, tol should generally be increased
-#' @return a list 
-#' @return \item{u}{matrix of size (number of regressors * number of components), contains the component-loadings, 
-#' i.e. the coefficients of the regressors in the linear combination giving each component}
-#' @return \item{comp}{matrix of size (number of statistical units * number of components) having the components as column vectors}
-#' @return \item{compr}{matrix of size (number of statistical units * number of components) having the standardized components as column vectors}
-#' @return \item{ds}{the final value of the regularization degree}
+# @title Estimation algorithm for K components
+# @description calculates the K components by iteratively calling function oneComponent 
+# @export
+# @param X matrix (n*p) containing the standardized covariates
+# @param Y matrix (n*q) containing dependent variables
+# @param AX matrix of additional covariates used in the generalized regression but not entering the linear 
+# combinations giving components
+# @param K integer specifying the number of components
+# @param family a vector of the same length as the number of responses containing characters 
+# identifying the distribution families of the dependent variables.
+# "bernoulli", "binomial", "poisson" or "gaussian" are allowed.
+# @param size matrix of size statistical units * number of binomial responses, giving the number of trials 
+# for binomial dependent variables. 
+# @param offset used for the poisson dependent variables.
+# A vector or a matrix of size: number of observations * number of Poisson dependent variables is expected
+# @param crit crit a list of two elements : maxit and tol, describing respectively the maximum number of iterations and 
+# the tolerance convergence criterion for the Fisher scoring algorithm. Default is set to 50 and 10e-6 respectively. 
+# @return a list 
+# @return \item{u}{matrix of size (number of regressors * number of components), contains the component-loadings, 
+# i.e. the coefficients of the regressors in the linear combination giving each component}
+# @return \item{comp}{matrix of size (number of statistical units * number of components) having the components as column vectors}
+# @return \item{compr}{matrix of size (number of statistical units * number of components) having the standardized components as column vectors}
+# @return \item{ds}{the final value of the regularization degree}
 kComponents <- function(X,Y,AX,K,family,size=NULL,offset=NULL,crit=list())
 {
   n<-dim(X)[1]

R/metric.r

@@ -1,6 +1,6 @@
-#'@title Function that computes a submatrix of metric M^(-1/2)
-#'@export 
-#'@param x the column vector of a regressor or factor
+#@title Function that computes a submatrix of metric M^(-1/2)
+#@export 
+#@param x the column vector of a regressor or factor
 metricBloc <- function(x){
   z <- model.matrix(~x)[,-1,drop=FALSE]
   z <- scale(z,center=TRUE,scale=FALSE)
@@ -9,9 +9,9 @@ metricBloc <- function(x){
 }
 
 
-#'@title Function that computes M^(-1/2)
-#'@export 
-#'@param data data.frame containing all covariates
+#@title Function that computes M^(-1/2)
+#@export 
+#@param data data.frame containing all covariates
 metric <- function(data){
   z <- sapply(data,metricBloc)
   z <- bdiag(z)

R/multivariateFormula.r

@@ -1,17 +1,17 @@
 #' @title Formula construction
-#' @description Construction of the multivariate scglr formula
+#' @description Construction of the multivariate scglr formula.
 #' @export
-#' @param namesY a vector of character containing the names of the dependent variables
+#' @param namesY a vector of character containing the names of the dependent variables.
 #' @param namesX a vector of character containing the names of the covariates (X) involved
-#'  in the components
-#' @param namesSX a vector of character containing the names of the additional covariates
-#' @return an object of class formula
-multivariateFormula <- function(namesY,namesX,namesSX=NULL)
+#'  in the components.
+#' @param namesAX a vector of character containing the names of the additional covariates.
+#' @return an object of class \code{Formula}.
+multivariateFormula <- function(namesY,namesX,namesAX=NULL)
 {
   forml <- paste(namesY,collapse="+")
   formr <- paste(namesX,collapse="+")
-  if(!is.null(namesSX)){
-   formr2 <- paste(namesSX,collapse="+")
+  if(!is.null(namesAX)){
+   formr2 <- paste(namesAX,collapse="+")
    form <- paste(paste(forml,"~",formr,sep=""),formr2,sep="|")
    formula <- as.Formula(form)
   }else{