boundary {exactmaxsel}R Documentation

Computes the coordinates of the boundaries

Description

The function boundary is an internal function that computes the greatest (and also the smallest if lower=TRUE) number of observations from class Y=1 in the left node that lead to an association criterion <= c. Remark: in general, the numbers output by boundary are not integers.

Usage

boundary(x, n0, n1, c, statistic, lower=TRUE)

Arguments

x the number of observations in the left node
n0 the number of observations in class Y=0
n1 the number of observations in class Y=1
c the value of the criterion that should not be exceeded.
statistic the association measure. Currently, only statistic="chi2" (chi-square statistic) and statistic="gini" (the Gini-gain from machine learning) are implemented.
lower Should the lower boundary also be computed?

Details

This function should not be called by the user in practice.

Value

a list with

upper the upper boundary (greatest allowed value).
lower the lower boundary (smallest allowed value).

Author(s)

Anne-Laure Boulesteix (http://www.slcmsr.net/boulesteix).

References

A.-L. Boulesteix (2006), Maximally selected chi-square statistics for ordinal variables, Biometrical Journal 48:451-462.

A.-L. Boulesteix (2006), Maximally selected chi-square statistics and binary splits of nominal variables, Biometrical Journal 48:838-848.

C. Strobl, A.-L. Boulesteix and T. Augustin (2007), Unbiased split selection for classification trees based on the Gini index, Computational Statistics and Data Analysis (in press).

See Also

Ford, Fcat.

Examples

# load exactmaxsel library
library(exactmaxsel)

boundary(10,30,30,c=3,statistic="chi2",lower=TRUE)


[Package exactmaxsel version 1.0-2 Index]