stable: Classification Stability

In vegsoup: Classes and Methods for Phytosociology

Description

Adjusted Goodman-Kruskal's lambda index of classification stability calculated from cross-tabulation of partitions (clusters) obtained from the data set and the subset selected by without-replacement bootstrap (Tichý et al. 2011). Assignments are evaluated by Goodman-Kruskal’s lambda index and a modified version of lambda.

Usage

## S4 method for signature 'VegsoupPartition'
stable(x, nitr = 200, nitr.lambda = 10, ...)

Arguments

x	VegsoupPartition object.
nitr	number of iterations.
nitr.lambda	number of iteration to calculate random values of lambda.
...	arguments passed to VegsoupPartition except argument method which is obtained from the object x.

Value

returns a list with components

lambda	lambda statistic, values range from 0 (no association) to 1 (the theoretical maximum possible association)
modified.lambda	modified lambda statistic

Note

This function is not fully tested.

Author(s)

adapted by Kaiser based on R-code supplied with the JUICE software

References

Tichý, L. and Chytrý, M. and Šmarda, P. (2011) Evaluating the stability of the classification of community data. Ecography 34:807-813

See Also

OptimStride, sample

Examples

x <- VegsoupPartition(coenoflex(seed = 1234), k = 3)

stable(x, nitr = 19, nitr.lambda = 10)

vegsoup documentation built on Feb. 24, 2021, 3 a.m.