optimize {vegsoup} | R Documentation |
Modify the partitioning of an object as to optimize silhouette widths (distance based method) or Dufrene and Legendre's indicator value (original data).
## S4 method for signature 'VegsoupPartition' optsil(x, maxitr = 100, verbose = FALSE, ...) ## S4 method for signature 'VegsoupPartition' optindval(x, maxitr = 100, minsiz = 5, verbose = FALSE, ...)
x |
|
maxitr |
integer. The maximum number of iterations to perform. |
verbose |
logical. Print elapsed CPU time. |
minsiz |
integer. The minimum size of the partition to consider reassigning a sample out of. |
... |
arguments passed to |
optsil
is a simple wrapper for function optsil
in package optpart to polish an existing clustering.
optindval
interfaces function optindval
in package optpart. This method maximizes the Dufrene and Legendre's indicator value (Indval).
Returns a modified object of class 'VegsoupPartition'
with possibly changed partitioning vector.
Both methods can be very slow when applied to big data sets! optsil
is usually faster.
Roland Kaiser, implemented for vegsoup using Dave W. Roberts optsil
and optindval
procedures of package optpart.
Dufrene, M. and Legendre, P. 1997 Species assemblages and indicator species: The need for a flexible asymmetrical approach. Ecological Monographs, 67,345–366.
Rousseeuw, P. 1987 Silhouettes: A graphical aid to the interpretation and validation of cluster analysis. Journal of Computational and Applied Mathematics, 53,53–65.
Roberts, D. 2015 Vegetation classification by two new iterative reallocation optimization algorithms. Plant Ecology, 216,741–758.
require(vegsoup) data(windsfeld) x <- VegsoupPartition(windsfeld, k = 5) ## Not run: ## wait until optpart registers S3methods xs <- optsil(x, verbose = TRUE) xi <- optindval(x, verbose = TRUE) confusion(xs, xi) ## End(Not run)