% building this document: (in R) Sweave ("ctc.Rnw")
\documentclass[a4paper]{article}

\title{Amap Package}
\author{Antoine Lucas}
%\VignetteIndexEntry{Introduction to amap}
%\VignettePackage{amap}
\SweaveOpts{echo=FALSE}
%\usepackage{a4wide}

\begin{document}

\maketitle

\tableofcontents

\section{Overview}

{\tt Amap} package includes standard hierarchical  
clustering and k-means. We optimize implementation 
(with a parallelized hierarchical clustering) and
allow the possibility of using different distances like
Eulidean or Spearman (rank-based metric).

We implement a principal component analysis (with robusts methods).

\section{Usage}

\subsection{Clustering}

The standard way of building  a hierarchical clustering:
<<echo=TRUE>>=
library(amap)
data(USArrests)
h = hcluster(USArrests)
plot(h)
@ 
 Or for the ``heatmap'':
<<echo=TRUE>>=
heatmap(as.matrix(USArrests),
        hclustfun=hcluster,
        distfun=function(u){u})
@ 
On a multiprocessor computer:
<<echo=TRUE>>=
h = hcluster(USArrests,nbproc=4)   
@ 
The K-means clustering:
<<echo=TRUE,results=hide>>=
Kmeans(USArrests,centers=3,method="correlation")
@ 

\subsection{Robust tools}

A robust variance computation:
<<echo=TRUE,results=hide>>=
data(lubisch)
lubisch <- lubisch[,-c(1,8)]
varrob(scale(lubisch),h=1)
@ 
A robust principal component analysis:
<<echo=TRUE,results=hide>>=
p <- acpgen(lubisch,h1=1,h2=1/sqrt(2))
plot(p)
@ 
Another robust pca:
<<echo=TRUE,results=hide>>=
p <- acprob(lubisch,h=4)
plot(p)
@ 


\section{See Also}

Theses examples can be tested with command
{\tt demo(amap)}.\\

\noindent
All functions has got man pages, try 
{\tt help.start()}.\\

\noindent
Robust tools has been published: \cite{caussinu+ruiz} and
\cite{caussinu+ruiz2}.


\bibliographystyle{plain}
\bibliography{amap}



\end{document}