version 2.0.3

DESCRIPTION

@@ -1,27 +1,30 @@
 Package: SCGLR
 Type: Package
 Title: Supervised Component Generalized Linear Regression
-Version: 2.0.2
-Date: 2014-02-21
+Version: 2.0.3
+Date: 2016-03-07
 Authors@R: c(
   person("Guillaume", "Cornu", role=c("cre","aut"),email="gcornu@cirad.fr"),
   person("Frederic", "Mortier", role="aut", email="fmortier@cirad.fr"),
   person("Catherine", "Trottier", role="aut", email="catherine.trottier@univ-montp3.fr"),
   person("Xavier", "Bry", role="aut", email="xavier.bry@univ-montp2.fr"))
-Description: The Fisher Scoring Algorithm is extended so as to combine Partial Least Squares
-    regression with Generalized Linear Model estimation in the multivariate context.
+Description: The Fisher Scoring Algorithm is extended so as to combine Partial
+    Least Squares regression with Generalized Linear Model estimation in the
+    multivariate context.
 License: CeCILL-2 | GPL-2
 Depends: R (>= 3.0.0)
-Imports: Matrix,Formula,expm,graphics,ggplot2,grid,pROC
+Imports: Matrix,Formula,expm,graphics,grid,ggplot2 (>=
+        2.0.0),pROC,scales
 Suggests: parallel,gridExtra,knitr
 LazyData: yes
 VignetteBuilder: knitr
-Packaged: 2015-02-16 15:21:24 UTC; cornu
+RoxygenNote: 5.0.1
+NeedsCompilation: no
+Packaged: 2016-03-16 15:36:12 UTC; cornu
 Author: Guillaume Cornu [cre, aut],
   Frederic Mortier [aut],
   Catherine Trottier [aut],
   Xavier Bry [aut]
 Maintainer: Guillaume Cornu <gcornu@cirad.fr>
-NeedsCompilation: no
 Repository: CRAN
-Date/Publication: 2015-02-16 17:45:46
+Date/Publication: 2016-03-16 18:16:10

MD5

@@ -1,56 +1,56 @@
-f9bcb4976f157950ef97c4b1f226e687 *DESCRIPTION
-dbde8a37f4c7fb6a9efecd3f12287875 *NAMESPACE
-b89bab137bd1feaad34216cd2df44180 *NEWS
+db31a13cecd4b6d470a1befe62a75351 *DESCRIPTION
+4ba9c634508ab826c4382c5afaa44e73 *NAMESPACE
+b14af99c4d1645c166b7bcc169884a74 *NEWS
 fe407b16523faae1e21b97df7b7602ce *R/critConvergence.r
 f7270f67b2a5ec01b786ce271a351dff *R/customize.r
 c769fb69d1872f5bd808b5c35d5692a9 *R/data.r
 f4253e5a2d3920e298ce31e34d0c4fe5 *R/f2x.r
-a646dc9a8fb3aa5e6b8160b265e59fca *R/infoCriterion.r
+2c64c6e90446cd513ea941d2382a6f8b *R/infoCriterion.r
 3dbdb5168b05e9ec4e0da2776475ec4d *R/ingUtils.r
 0349b37413d57e6d921c8d43efcc7eee *R/kComponents.r
 4ba7230033c3a7490d0dd9cc79e2bd15 *R/methods.r
-7bd55b5f836056a7a58f73375bf2d055 *R/metric.r
+e1c5c0463afa92947f89b4bc9d45b53b *R/metric.r
 e15ec931f554e11541a96ff4ee1e99f5 *R/multivariateFormula.r
-26ff07ca0c50f66a4007594d68b2f46b *R/multivariateGlm.fit.r
-a41599c03c0d126edecb0dca6bbd7a9d *R/multivariateGlm.r
-476413c3061140b69bae791da7bc28c9 *R/multivariatePredictGlm.r
+06aec9e8efcba1d2329a42797f7167e9 *R/multivariateGlm.fit.r
+22771fc611fbe00de6de88f2ea8effa9 *R/multivariateGlm.r
+ee5f8640fae5c3618002c4df0fb2404b *R/multivariatePredictGlm.r
 658856a3e97ad270dcec6757f531fe5d *R/oneComponent.r
-af3562bfd9a5f7d37359be6780899f5c *R/package.r
-40bd86d9973e18d9cf60fd32fca625c8 *R/plot.r
+ce526cb03a60470bb399ad4db6d8c89d *R/package.r
+c28e5a9b122be63d02e32467fafc603b *R/plot.r
 6b6c89a00bd175b725611565e868f1dd *R/print.r
-f0e3d52b7a0e6868e7ee81991ac06c8c *R/scglr.r
-fb8926305414a78135bf96b4f4a59e22 *R/scglrCrossVal.r
-3d486ffb933f86ee95ff48460220f5f1 *R/summary.r
+095ccd6b8d8b9434f8a71fbd7c3d52d9 *R/scglr.r
+0e536fcc9b4828613ab2f6a136279204 *R/scglrCrossVal.r
+7f0cf1c70134e4606b0b8be533447965 *R/summary.r
 6134b8018a01f25abdacad8439433b4b *R/utils.r
-a1baa945407de24483e5f6c5b0770ea4 *build/vignette.rds
+f2fcc552a6233f3301847a057141801a *build/vignette.rds
 b454e91b945abf3eaaa1d2ddf9c52cfc *data/genus.rda
 7228f976812aee4fdd60f7ac760b4a43 *demo/00Index
 7ff42f88c9d37c33feb1dea01efbeebe *demo/scglr.r
 725274da96c067732984971ed539f43d *demo/scglrCrossVal.r
 619f968ac2ad86861dcb639dcc703aa8 *demo/scglrPlots.r
 bb881a7cd25753a10fb2af5b7b85beed *demo/scglrPrediction.r
-c6e040e085bdd5205853378e106275d0 *inst/doc/scglrVignettes.R
-2682315ff83e302194912d8c2fca551c *inst/doc/scglrVignettes.pdf
+653cab14380859c07bed99b532c29f13 *inst/doc/scglrVignettes.R
+4fb8b6c0594a58ade3a08a6696551e50 *inst/doc/scglrVignettes.pdf
 0e57f56782f8048a2cf81a46c20af774 *inst/doc/scglrVignettes.rnw
 725274da96c067732984971ed539f43d *inst/examples/exCrossVal.r
 619f968ac2ad86861dcb639dcc703aa8 *inst/examples/exPlots.r
 bb881a7cd25753a10fb2af5b7b85beed *inst/examples/exPrediction.r
 7ff42f88c9d37c33feb1dea01efbeebe *inst/examples/exScglr.r
-41b723814971ea3ae23c72e018592966 *man/barplot.SCGLR.Rd
-942169004fe3d29b13b2aaa5254c059f *man/critConvergence.Rd
-337143294908e9c7f41d07c76d1111f8 *man/customize.Rd
-fae23e61ce55d56fd95088e3b002f572 *man/genus.Rd
-885860d1502f83b3e5fa3fbccfe394cf *man/infoCriterion.Rd
-d2d4758e6975a58f4bf0bfd12001edfd *man/method.Rd
-e16c4519eac5481c61cb9593ca4f9290 *man/multivariateFormula.Rd
-251425b7f53977e906245e2888d9cd87 *man/multivariateGlm.Rd
-eb969fdb8b86ea9f364980dff2ce9fdc *man/multivariatePredictGlm.Rd
-2e0383d57e3e9eba920110ba82172200 *man/pairs.SCGLR.Rd
-56683bb246ae374589f4be164f91d254 *man/plot.SCGLR.Rd
-b77acaeb5365ae2ae22568903f20c856 *man/print.SCGLR.Rd
-63a59c5dff75029689a25dcae05ee672 *man/scglr-package.Rd
-cf0386eb71df5bec37d842c3358a5dc3 *man/scglr.Rd
-99bd0d94029e9d7d6001a1df89f57389 *man/scglrCrossVal.Rd
-64a6acc2dfca6474d1978b4079eaf254 *man/summary.SCGLR.Rd
+206778dfb44c78863d4c4d9378aa59ac *man/barplot.SCGLR.Rd
+89b10c634cb3dccab7f0762c6c4f3cdd *man/critConvergence.Rd
+65200f77b0a333685763c238aa357127 *man/customize.Rd
+7461eef8fad93cd4ac157377ccf17192 *man/genus.Rd
+47a337edb89a436f4a1028c3c993c43a *man/infoCriterion.Rd
+23fea70f5b37cd428f5ef91ac344e863 *man/method.Rd
+43c08f871673d7f1d661d8bc61507cf5 *man/multivariateFormula.Rd
+5f1c2fdafac5921930a8e4da5222e462 *man/multivariateGlm.Rd
+fb15281d171481ad3a0feb5c11a7b9a1 *man/multivariatePredictGlm.Rd
+d867276eec5061ce9db9ceec190a98f5 *man/pairs.SCGLR.Rd
+b1aff9203758ed32dd92ac307800c457 *man/plot.SCGLR.Rd
+c59b7c2b1a1fa68d474ce4e19b2a6779 *man/print.SCGLR.Rd
+41147a0a7a4d4ed29be57376c19947e9 *man/scglr-package.Rd
+a745d02985c5e7130af05af1bba62976 *man/scglr.Rd
+31dfe9c98b683b0ab27bb69b2728a287 *man/scglrCrossVal.Rd
+aff812e2db18dc00f89e3404d9905ba8 *man/summary.SCGLR.Rd
 3088a2c384aac3c5287caca92e3cbf07 *vignettes/bibScglr.bib
 0e57f56782f8048a2cf81a46c20af774 *vignettes/scglrVignettes.rnw

NAMESPACE

@@ -1,4 +1,4 @@
-# Generated by roxygen2 (4.1.0): do not edit by hand
+# Generated by roxygen2: do not edit by hand
 
 S3method(barplot,SCGLR)
 S3method(pairs,SCGLR)
@@ -19,7 +19,24 @@ export(scglrCrossVal)
 import(Formula)
 import(expm)
 import(ggplot2)
-import(grid)
 import(pROC)
 importFrom(Matrix,bdiag)
 importFrom(graphics,barplot)
+importFrom(grid,circleGrob)
+importFrom(grid,gpar)
+importFrom(grid,grid.layout)
+importFrom(grid,grid.newpage)
+importFrom(grid,pushViewport)
+importFrom(grid,viewport)
+importFrom(scales,percent_format)
+importFrom(stats,aggregate)
+importFrom(stats,as.formula)
+importFrom(stats,coef)
+importFrom(stats,cor)
+importFrom(stats,dbinom)
+importFrom(stats,dnorm)
+importFrom(stats,dpois)
+importFrom(stats,glm)
+importFrom(stats,model.extract)
+importFrom(stats,model.matrix)
+importFrom(utils,combn)

NEWS

@@ -1,3 +1,8 @@
+* version 2.0.3
+- added utils::globalVariables and @importFrom to keep CMD CHECK happy
+- added plane parameter to barplot to have alternate coloring for selected components
+- now require ggplot2 >= 2.0.0
+
 * version 2.0.2
 - fix bug in component optimization
 

R/infoCriterion.r

@@ -1,5 +1,7 @@
 #' @title Function that calculates cross-validation selection criteria
 #' @export
+#' @importFrom stats dbinom dpois dnorm
+#' @description Function that calculates cross-validation selection criteria
 #' @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 

R/metric.r

@@ -1,6 +1,7 @@
 #@title Function that computes a submatrix of metric M^(-1/2)
 #@export 
 #@param x the column vector of a regressor or factor
+#' @importFrom stats model.matrix
 metricBloc <- function(x){
   z <- model.matrix(~x)[,-1,drop=FALSE]
   z <- scale(z,center=TRUE,scale=FALSE)

R/multivariateGlm.fit.r

@@ -1,5 +1,6 @@
 #' @rdname multivariateGlm
 #' @export 
+#' @importFrom stats as.formula glm 
 #' @param Y matrix of dependent variables.
 #' @param comp matrix of covariates.
 # @param family a vector of character giving the family distribution of each response

R/multivariateGlm.r

@@ -2,7 +2,8 @@
 #' @description \code{multivariateGlm} is used to fit multivariate generalized linear models
 #' specified by a symbolic formula together with the distributions of the responses. 
 #' This function performs a simple GLM fit for each dependent variable with the associated distribution.
-#' @export 
+#' @export
+#' @importFrom stats model.matrix model.extract 
 #' @param formula an object of class \code{Formula} (or one that can be coerced to that class): a symbolic description of the model to be fitted.
 #' @param data the data frame to be modeled.
 #' @param family a vector of character giving the family distribution of each response.

R/multivariatePredictGlm.r

@@ -1,5 +1,7 @@
 #' @title Function that predicts the responses from the covariates for a new sample
 #' @export
+#' @importFrom stats as.formula model.matrix
+#' @description  Function that predicts the responses from the covariates for a new sample
 #' @param Xnew a data frame containing the values of the covariates for the new sample.
 #' @param family a vector of character specifying the distributions of the responses.
 #' @param beta the matrix of coefficients estimated from the calibration sample.

R/package.r

@@ -16,7 +16,7 @@
 #' @import Formula
 #' @import expm
 #' @importFrom graphics barplot
-#' @import ggplot2 grid
+#' @import ggplot2
 #' @import pROC
 NULL
 

R/plot.r

@@ -1,9 +1,15 @@
-# remove warnings due to ggplot2 syntax strangeness
-utils::globalVariables(c("comp","y","label","angle","hjust"))
+if(getRversion()>="2.15.1") {
+  # remove warnings due to ggplot2 syntax strangeness
+  utils::globalVariables(c("comp","y","label","angle","hjust"))
+}
+
 
 #' SCGLR generic plot
 #' @export
+#' @importFrom stats cor aggregate
+#' @importFrom grid circleGrob gpar
 #' @method plot SCGLR
+#' @description SCGLR generic plot
 #' @param x an object from SCGLR class.
 #' @param \dots optional arguments (see \link{customize}).
 #' @param style named list of values used to customize the plot (see \link{customize})
@@ -150,14 +156,21 @@ plot.SCGLR <- function(x, ..., style=getOption("plot.SCGLR"), plane=c(1,2)) {
   # build base plot
   p <- qplot((-1:1)*cust("expand",1.0), (-1:1)*cust("expand",1.0), geom="blank")+
     coord_fixed()+
-    # thicker x unit arrow
     xlab(paste(axis_names[1],"(",round(100*inertia[1],2),"%)")) + 
-    geom_hline(yintercept=0)+
-    geom_segment(aes(x=-1.1,xend=1.1,y=0,yend=0),size=1,arrow=arrow(length=unit(0.02,"npc")))+
+    ylab(paste(axis_names[2],"(",round(100*inertia[2],2),"%)")) 
+
+  if(observations) {
+    p <- p +
+      geom_hline(yintercept=0,size=0.5)+
+      geom_vline(xintercept=0,size=0.5)
+  }
+  if(covariates||predictors) {
+    p <- p +
+    # thicker x unit arrow
+      geom_segment(aes(x=-1.1,xend=1.1,y=0,yend=0),size=1,arrow=arrow(length=unit(0.02,"npc")))+
     # thicker y unit arrow
-    ylab(paste(axis_names[2],"(",round(100*inertia[2],2),"%)")) + 
-    geom_vline(xintercept=0)+
-    geom_segment(aes(y=-1.1,yend=1.1,x=0,xend=0),size=1,arrow=arrow(length=unit(0.02,"npc")))
+      geom_segment(aes(y=-1.1,yend=1.1,x=0,xend=0),size=1,arrow=arrow(length=unit(0.02,"npc")))
+  }
 
   # plot title
   if(has_cust("title", FALSE)) {
@@ -209,6 +222,8 @@ plot.SCGLR <- function(x, ..., style=getOption("plot.SCGLR"), plane=c(1,2)) {
         alpha=cust("observations.alpha",1)
       )
     }
+    tmp <- data.frame(x=c(0,0),y=range(obs$y))
+    p <- p + geom_line(aes(x,y),data=tmp)
   }
 
   # add linear predictor arrows
@@ -364,7 +379,7 @@ plot.SCGLR <- function(x, ..., style=getOption("plot.SCGLR"), plane=c(1,2)) {
 
     # draw labels ?
     if(cust("factor.labels",TRUE)) {
-      p <- p + geom_text(
+      p <- p + geom_label(
         aes_string(x="x",y="y",label=factor),
         data=bary,
         color=cust("factor.labels.color","black"),
@@ -380,20 +395,52 @@ plot.SCGLR <- function(x, ..., style=getOption("plot.SCGLR"), plane=c(1,2)) {
 #' @title Barplot of percent of overall X variance captured by component
 #' @export
 #' @method barplot SCGLR
+#' @description A custom plot for SCGLR objetcs
 #' @param height object of class 'SCGLR', usually a result of running \code{\link{scglr}}.
 #' @param \dots optional arguments.
+#' @param plane a size-2 vector (or comma separated string) indicating which components are plotted (eg: c(1,2) or "1,2").
 #' @return an object of class ggplot.
 #' @seealso For barplot application see examples in \code{\link{plot.SCGLR}}.
-barplot.SCGLR <- function(height, ...) {
-  inertia <- data.frame(inertia=height$inertia,comp=1:length(height$inertia))
-  qplot(comp, inertia, data=inertia, geom="bar", stat="identity", width=0.5) +
+#' @importFrom scales percent_format
+barplot.SCGLR <- function(height, ..., plane=NULL) {
+  if(!is.null(plane)) {
+    # process plane
+    if(is.character(plane)) {
+      plane <- as.integer(trim(unlist(strsplit(plane, ","))))
+    }
+
+    # sanity checking
+    if(length(plane) !=2 ) {
+      stop("Plane should have two components!")
+    }
+    if((min(plane)<1) || (max(plane)>length(height$inertia))) {
+      stop("Invalid components for plane!")
+    }
+  }
+
+  # build data frame from inertia
+  inertia <- data.frame(
+      inertia=height$inertia,
+      comp=1:length(height$inertia),
+      plane_color="black",
+      stringsAsFactors = F)
+
+  # recolor unused component in gray
+  if(!is.null(plane))
+    inertia$plane[-plane] <- "gray"
+
+  ggplot(data=inertia)+geom_bar(aes(comp,inertia),fill=inertia$plane,stat="identity",width=0.5) +
     scale_x_discrete(labels=names(height$inertia),limits=1:length(inertia$comp))+
-    labs(x="Components", y="Inertia", title="Inertia per component\n", ...)
+    scale_y_continuous(labels=percent_format())+
+    labs(x="Components", y="Inertia", title="Inertia per component\n", ...)#+
+    #geom_text(aes(comp,inertia,label=round(100*inertia,2)),vjust=0)
 }
 
 #' @title Pairwise scglr plot on components
 #' @export
+#' @importFrom utils combn
 #' @method pairs SCGLR
+#' @description Pairwise scglr plot on components
 #' @param x object of class 'SCGLR', usually a result of running \code{\link{scglr}}.
 #' @param \dots optionally, further arguments forwarded to \code{link{plot.SCGLR}}.
 #' @param nrow number of rows of the grid layout.
@@ -441,6 +488,7 @@ pairs.SCGLR <- function(x, ..., nrow=NULL, ncol=NULL, components=NULL) {
 }
 
 ## equivalent du par pour les ggplot2
+#' @importFrom grid viewport grid.newpage pushViewport viewport grid.layout
 vp.layout <- function(x, y) viewport(layout.pos.row=x, layout.pos.col=y)
 arrange <- function(..., nrow=NULL, ncol=NULL, as.table=FALSE) {
   dots <- list(...)

R/scglr.r

@@ -1,6 +1,11 @@
+if(getRversion()>="2.15.1") {
+  utils::globalVariables(c("na.omit","coef"))
+}
+
 #' @title Function that fits the scglr model
 #' @description Calculates the components to predict all the dependent variables.
 #' @export scglr
+#' @importFrom stats model.matrix model.extract coef cor
 #' @param formula an object of class \code{Formula} (or one that can be coerced to that class): a symbolic description of the model to be fitted.
 #' @param data  a data frame to be modeled.
 #' @param family a vector of character of the same length as the number of dependent variables: 

R/scglrCrossVal.r

@@ -1,5 +1,10 @@
+if(getRversion()>="2.15.1") {
+  utils::globalVariables(c("na.omit","coef"))
+}
+
 #' Function that fits and selects the number of component by cross-validation.
 #' @export 
+#' @importFrom stats model.matrix model.extract coef cor
 #' @param formula an object of class "Formula" (or one that can be coerced to that class): a symbolic description of the model to be fitted.
 #' @param data the data frame to be modeled.
 #' @param family a vector of character of length q specifying the distributions of the responses. Bernoulli, binomial, poisson and gaussian are allowed.

R/summary.r

@@ -1,4 +1,6 @@
 #' @export
+#' @importFrom stats cor
+#' @importFrom utils combn
 #' @title Summarizing SCGLR fits
 #' @description Summary method for class "SCGLR".
 #' @method summary SCGLR