Calculates a CART histogram

Share:

Description

Calculates a CART histogram. The estimate is represented as an evaluation tree. An CART histogram is a multivariate adaptive histogram which is obtained by pruning an evaluation tree of an overfitting histogram.

Usage

1
eval.cart(dendat, leaf, minobs = NULL)

Arguments

dendat

n*d data matrix

leaf

positive integer; the cardinality of the partition of the histogram

minobs

non-negative integer; splitting of a bin of the overfitting histogram will be continued if the bin containes "minobs" or more observations

Details

The partition of the histogram may not contain exactly "leaf" rectangles: the cardinality of the partition is as close as possible to "leaf"

Value

An evaluation tree

Author(s)

Jussi Klemela

See Also

lstseq.cart, densplit

Examples

1
2
3
4
5
6
library(denpro)
dendat<-sim.data(n=600,seed=5,type="mulmodII")
eva<-eval.cart(dendat,16)

dp<-draw.pcf(eva,pnum=c(60,60))
persp(dp$x,dp$y,dp$z,theta=-20,phi=30)