mappeR: Main mappeR function
In arnhew99/mappeR: Implements a topological visualisation algorithm

Description Usage Arguments Details Value Author(s) References Examples

Runs the mappeR algorithm on a data set with the given parameters. The output is an adjacency matrix that specifies the edges between nodes (suitable for use with packages like igraph).

1	mappeR(X, n.slices, overlap, filterfn, clusterfn)

`X`	Data matrix
`n.slices`	Number of overlapping subsets to slice the data into
`overlap`	Amount of overlap between clusters (0-50)
`filterfn`	Function to filter data
`clusterfn`	Function to cluster inside subsets

A filter function should take as argument X and return a vector of length equal to the number of observations (rows) of X.
A cluster function should take as argument X (internally to mappeR, this will be the data in a slice), and return a list of cluster members.

mappeR returns an object of class "mapper".

An object of class "mapper" is an environment containing at least the following components:

`adjmat`	Node adjacency matrix
`filter.values`	Vector of filter values taken by each observation in X
`clusters`	List of observations in each cluster
`slices`	Vector of number of clusters in each slice

Matthew Arnold

Singh, G., Memoli, F., Carlsson, G. (2007) Topological Methods for the Analysis of High Dimensional Data Sets and 3D Object Recognition

# compare output to pairs plot of iris data
pairs(iris[,1:4], main = "Anderson's Iris Data -- 3 species",
      pch = 21, bg = c("red", "green3", "blue")[unclass(iris$Species)])

iris.mappeR.out <- mappeR(iris[,1:4],6,50,filterEccentricity,singlelinkagehist)

# find the iris types in each cluster (node)
iris.names <- lapply(iris.mappeR.out$clusters, function(x) return(substr(paste(unique(iris[x,5])),1,2)))
cluster.names <- sapply(iris.names, paste, collapse="/")

# plotting is done with igraph using the adjacency matrix in the mappeR output 
plot(iris.mappeR.out,vertex.label=cluster.names,vertex.size=30,edge.width=2)

# identify a circle
x <- runif(2000,0,2*pi)
X <- cbind(cos(x),sin(x))
plot(mappeR(X,6,50,filterCol1,singlelinkagehist))