Check results for 'gbev'

using R Under development (unstable) (2012-05-19 r59366)
using platform: x86_64-unknown-linux-gnu (64-bit)
using session charset: UTF-8
checking for file ‘gbev/DESCRIPTION’ ... OK
this is package ‘gbev’ version ‘0.1.1’
checking package dependencies ... OK
checking if this is a source package ... OK
checking if there is a namespace ... WARNING
As from R 2.14.0 all packages need a namespace.
One will be generated on installation, but it is better to handcraft a
NAMESPACE file: R CMD build will produce a suitable starting point.
CRAN requires a NAMESPACE file for all submissions.
checking for executable files ... OK
checking whether package ‘gbev’ can be installed ... OK
checking installed package size ... OK
checking package directory ... OK
checking for portable file names ... OK
checking for sufficient/correct file permissions ... OK
checking DESCRIPTION meta-information ... OK
checking top-level files ... OK
checking index information ... OK
checking package subdirectories ... OK
checking R files for non-ASCII characters ... OK
checking R files for syntax errors ... OK
checking whether the package can be loaded ... OK
checking whether the package can be loaded with stated dependencies ... OK
checking whether the package can be unloaded cleanly ... OK
checking loading without being on the library search path ... OK
checking for unstated dependencies in R code ... OK
checking S3 generic/method consistency ... OK
checking replacement functions ... OK
checking foreign function calls ... OK
checking R code for possible problems ... NOTE
File ‘gbev/R/gbev.R’:
.First.lib calls:
require(mvtnorm)

Package startup functions should not change the search path.
See section ‘Good practice’ in ?.onAttach.

gbev: warning in match.call(expand = FALSE): partial argument match of
'expand' to 'expand.dots'
gbev: no visible binding for global variable ‘response’
gbev: no visible binding for global variable ‘var.names’
part.dep: no visible binding for global variable ‘fit’
checking Rd files ... OK
checking Rd metadata ... OK
checking Rd cross-references ... OK
checking for missing documentation entries ... OK
checking for code/documentation mismatches ... OK
checking Rd \usage sections ... OK
checking Rd contents ... OK
checking for unstated dependencies in examples ... WARNING
‘library’ or ‘require’ call not declared from: ‘lattice’
checking line endings in C/C++/Fortran sources/headers ... OK
checking line endings in Makefiles ... OK
checking for portable use of $(BLAS_LIBS) and $(LAPACK_LIBS) ... OK
checking compiled code ... OK
checking examples ... ERROR
Running examples in ‘gbev-Ex.R’ failed
The error most likely occurred in:

> ### Name: gbev
> ### Title: Boosted regression trees with errors-in-variables
> ### Aliases: gbev
> ### Keywords: nonparametric tree
>
> ### ** Examples
>
>
> ### Univariate regression example
> n<-500
> varX<-1
> varME<-0.25
> varNoise<-0.3^2
>
> ### Data
> x<-rnorm(n,sd=sqrt(varX)) ### Error free covariate
> w<-x+rnorm(n,sd=sqrt(varME)) ### Error contaminated version
> fx<-sin(pi*x/2)/(1+2*(x^2)*((2*as.numeric(x>=0)-1)+1)) ### True regression function
> y<-fx+rnorm(n,sd=sqrt(varNoise)) ### Response
> dat<-data.frame(y=y,w=w)
>
> ### Measurement error model ####
> ###
> ### The measurement error model is a list of the following components:
> ###
> ### SigmaX: the covariance matrices of the mixture model for the error free covariates
> ### SigmaX[i,,] is the covariance matrix of the i-th mixture density
> ### mu: the means of the mixture model for the error free covariates
> ### mu[i,] is the mean-vector of the i-th mixture density
> ### SigmaME: the covariance matrix of the measurment error
> ### pComp: the weights of the mixture distribution, pComp[i] is the weight of the
> ### i-th mixture density
> ### numComp: the number of components in the mixture
> ##
> p<-1
> pME<-1
>
> numComp<-3 ## number of components in gaussian mixture for X-distribution
> SigmaME<-diag(varME,pME)
> SigmaJ<-array(dim=c(numComp,pME,pME))
> mu<-array(dim=c(numComp,pME))
> pComp<-array(1/numComp,dim=c(numComp,1))
> for(i in 1:numComp)
+ {
+ SigmaJ[i,,]<-diag(varX,pME)
+ mu[i,]<-rep(0,pME)
+ }
> ### list required by "gbev" for measurement error model
> meModel<-list(SigmaX=SigmaJ,mu=mu,SigmaME=SigmaME,pComp=pComp,numComp=numComp)
>
>
> fit<-gbev(y~w,data=dat,
+ measErrorModel=meModel,
+ method="L2", ## Squared error loss
+ nboost=1000, ## 1000 boosting iterations
+ lambda=5, ## regularization of regression tree
+ maxDepth=2, ## maximum tree depth, 2 corresponds stumps
+ mc=2, ## number of monte-carlo samples per tree build
+ minSplit=3, ## minimum number of obs in node to split
+ minBucket=0, ## minimum number of obs in nodes
+ sPoints=10, ## number of sampled candidate split points
+ intermPred=5) ## increments of iterations to store predictions
>
> ### 5-fold cross-validation
> hcv<-cvLoss(object=fit,k=5,random=FALSE,loss="L2")
> plot(hcv$iters,hcv$cvLoss,type="l")
>
> hp<-part.dep(object=fit,varIndx=1,firstTree=1,lastTree=hcv$estIter)
>
> x<-seq(-2,2,by=.02)
> fx<-sin(pi*x/2)/(1+2*(x^2)*((2*as.numeric(x>=0)-1)+1))
> points(x,fx,type="l",lty=5)
>
>
>
>
> ## Simulated binary regression example,
> ## with: Y=I( X1*X2+X2*X3+X1*X3>0), with measurement error on X's
> n<-1000
> p<-3
> varX<-1 ##
> varME<-0.5 ## measurement error variance
>
> x<-rnorm(p*n)
> x<-matrix(x,ncol=p,nrow=n)
> ## add measurement error
> w<-x+matrix(rnorm(p*n,sd=sqrt(varME)),ncol=p,nrow=n)
>
> x<-x[,c(1:p)]*x[,c(2:p,1)]
> x<-apply(x,1,sum)
> threshold<-0
> y<-as.numeric(x>threshold)
> dat<-data.frame(y=y,w1=w[,1],w2=w[,2],w3=w[,3]) ## must be modified if(p!=3)
>
>
> #### Measurement error model ######
> numComp<-1 ## Number of components in mixture
> SigmaME<-diag(varME,p) ## Covariance matrix of measurement error
> SigmaJ<-array(dim=c(numComp,p,p)) ## Covariance matices for mixture
> mu<-array(dim=c(numComp,p)) ## Mean vectors for mixture components
> pComp<-array(1/numComp,dim=c(numComp,1)) ## Mixture probabilities
> for(i in 1:numComp)
+ { ## filling in mixture model for X-distribution
+ SigmaJ[i,,]<-diag(varX,p)
+ mu[i,]<-rep(0,p)
+ }
> ## The list for measurement error model
> meModel<-list(SigmaX=SigmaJ,mu=mu,SigmaME=SigmaME,pComp=pComp,numComp=numComp)
>
> fit<-gbev(y~w1+w2+w3,data=dat,
+ measErrorModel=meModel,
+ method="logLike", ## loss function
+ nboost=1000, ## number of boosting iterations
+ lambda=40, ## regularization parameter used in regression tree
+ maxDepth=3, ## maximum depth of regression tree
+ minSplit=10, ## minimum number of observations in node to split
+ minBucket=0, ## minimum number in split node to allow split
+ sPoints=2, ## number of sampled canditate split points
+ mc=2, ## monte-carlo sample size used in each regression tree
+ intermPred=10) ## Increments of iterations to store loss function
>
>
> ## plot loss function as function of iterations
> hp<-plotLoss(fit,loss="logLike",startIter=10)
>
> ## bivariate partial dependence plot
> hdp<-part.dep(object=fit,varIndx=c(1,2),firstTree=1,
+ lastTree=1000,ngrid=50)
> dpp<-data.frame(x1=hdp$dat$x,x2=hdp$dat$y,prob=hdp$dat$z)
> library(lattice)
Error in library(lattice) : there is no package called ‘lattice’
Execution halted