prune.cubt {cubt}R Documentation

Prune a cubt tree.

Description

Prune a cubt tree.

Usage

 prune.cubt(trclust,Dt,mindist=NULL,qdist=0.8,alph=0.3,nleaves=NULL,dist= FALSE,...) 

Arguments

trclust

A CUBT tree.

Dt

Data used to construct the tree

mindist

Minimum distance between adjacent leaves to be pruned.

qdist

quantile of distances over couples of nodes in the tree to use as mindist.

alph

The quantile used for computing distances. See definition of the distance in he original paper.

nleaves

Number of leaves desired after pruning.

dist

If true returns the distances computed all along the nodes.

...

Other possible parameters passed for distance computation

Details

Minimum dissimilarity pruning. Two adjacent nodes are pruned if their distance defined by Fraiman et al. is less than mindist. Mindist may be specified by the user as a threshold, if not a quantile qdist is used, and mindist will be taken to be the corresponding quantile of the observed distances in the tree.

Examples

#See cubt help. 

[Package cubt version 3.3 Index]