%
% NOTE -- ONLY EDIT .Rnw!!!
% .tex file will get overwritten.
%
%\VignetteIndexEntry{@@PKNAME@@ Primer}
%\VignetteDepends{}
%\VignetteKeywords{Genomics}
%\VignettePackage{@@PKNAME@@}
%
% NOTE -- ONLY EDIT THE .Rnw FILE!!!  The .tex file is
% likely to be overwritten.
%
\documentclass[12pt]{article}

\usepackage{amsmath,pstricks}
\usepackage[authoryear,round]{natbib}
\usepackage{hyperref}


\textwidth=6.2in
\textheight=8.5in
%\parskip=.3cm
\oddsidemargin=.1in
\evensidemargin=.1in
\headheight=-.3in

\newcommand{\scscst}{\scriptscriptstyle}
\newcommand{\scst}{\scriptstyle}


\newcommand{\Rfunction}[1]{{\texttt{#1}}}
\newcommand{\Robject}[1]{{\texttt{#1}}}
\newcommand{\Rpackage}[1]{{\textit{#1}}}
\newcommand{\Rmethod}[1]{{\texttt{#1}}}
\newcommand{\Rfunarg}[1]{{\texttt{#1}}}
\newcommand{\Rclass}[1]{{\textit{#1}}}

\textwidth=6.2in

\bibliographystyle{plainnat} 
 
\begin{document}
%\setkeys{Gin}{width=0.55\textwidth}

\section{LogitBoost}

Boosting is a well-established approach to machine learning.
The basic idea is that predictions formed
by a series of `weak learners' are combined as cases are reweighted
to increase the weight assigned to `difficult' cases.

As implemented by Dettling and Buhlmann, the procedure
requires identification of a training set with predictors
and responses, and a test set for which responses should be
available.

The interface to logitboost created by Dettling is mildly inconvenient
for use with exprSets.  Records must be sorted
by response value.  The following function simplifies the
tasks of boosting with exprSets.  The second argument is
a character string naming the response variable of interest
(this will be converted to an integer in 0:nclasses).
Arguments to logitboost must be named and are passed through.
<<>>=
library(LogitBoost)
library(golubEsets)

lboostTwoEsets <- function(esetTrain, pdNameTrain,
     esetTest, ...) {
xlearn <- t(exprs(esetTrain))
ylearn <- as.numeric(factor(esetTrain[[pdNameTrain]])) - 1
ytest <- as.numeric(factor(esetTest[[pdNameTrain]])) - 1
orec <- order(ylearn)
xlearn <- xlearn[orec,]
ylearn <- ylearn[orec]
xtest <- t(exprs(esetTest))
logitboost(xlearn, ylearn, xtest, ...)
}
@
We now illustrate the procedure:
<<>>=
run1 <- lboostTwoEsets( golubTrain[20:2600,], "ALL.AML", 
 golubTest[20:2600,], mfinal=100 )
<<fig=TRUE>>=
print(summarize(run1, YTEST <<- as.numeric(factor(golubTest$ALL.AML))-1, mout=100))
@
<<>>=
table( apply(run1$probs,1,function(x)mean(x)>.5), YTEST )
@
%Modifications of interest include substitution of
%test for training set
 
@
\section{GBM}
<<>>=
 Y <- as.numeric(factor(golubTrain$ALL.AML))-1
 X <- t(exprs(golubTrain)[1:100,])
 dd <- data.frame(Y=Y, X)
library(gbm)

\end{document}