Optimum Classification by Counts of Indicator Species
Calculates the number of indicator species/cluster across a range of partitions
optimclass(comm, stride, pval = 0.01, counts = 2)
comm |
a community matrix with sample units as rows and species as columns |
stride |
an object of class ‘stride’from function
|
pval |
the minimum probability for inclusion in the list of indicators |
counts |
the minimum number of clusters for inclusion in the list |
Calculates the number of indicator species/cluster and the number of clusters with at least ‘counts’ indicators, using the φ index to identify indicators with probabilities less than or equal to ‘pval’. Arguably the optimal partition is the one with the most indicator species and the most clusters with adequate indicators.
A data.frame of
clusters |
number of clusters |
sig.spc |
the number of species with significant indicator value |
sig.clust |
the number of clusters with at least ‘counts’ indicator species |
The concept and first implementation were by Tichy in software package ‘Juice’, and this is a simple port of the algorithm to R.
Lubomir Tichy wrote the original algorithm
David W. Roberts droberts@montana.edu
Tichy, L., M. Chytry, M. Hajek, S. Talbot, and Z. Botta-Dukat. 2010. OptimClass: Using species-to-cluster fidelity to determine the optimal partition in classification of ecological communities. J. Veg. Sci. 21:287-299.
data(shoshveg) dis.bc <- dsvdis(shoshveg,'bray') opt.2.10 <- stride(2:20,dis.bc) ## Not run: optimclass(shoshveg,opt.2.10)
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.