scaspa: Finds the number of modes of histograms which are obtained by...

In delt: Estimation of Multivariate Densities Using Adaptive Partitions

Description

Function "densplit" returns an overfitting histogram as an evaluation tree. Function "prune" finds the candidate nodes for pruning. Function "scaspa" finds the number of modes (local maxima) in the histograms which are obtained by using these candidate nodes for pruning.

Usage

scaspa(treeseq, bind, eind)

Arguments

treeseq	a list returned by function "prune"
bind	integer in 1:subnum, where subnum is the number of elements in field "leafs" of "treeseq". We need bind<eind.
eind	integer in 1:subnum, where subnum is the number of elements in field "leafs" of "treeseq". We need bind<eind.

Value

List with the following vectors whose length is (eind-bind+1)

moodilkm	number of local maxima for each subtree
alfas	value of the smoothing parameter alpha for each subtree
leafnums	number of leaves for each subtree

Author(s)

Jussi Klemela

See Also

densplit, prune

Examples

set.seed(1)
dendat<-matrix(rnorm(20),10)  
minlkm<-2
et<-densplit(dendat,minlkm)
treeseq<-prune(et)
treeseq$leafs

scaspa(treeseq,1,5)

delt documentation built on May 2, 2019, 3:42 p.m.