diverclust: beta and gamma Diversity Clustering

In jarioksa/natto: An Extreme 'vegan' Package of Experimental Code

beta and gamma Diversity Clustering

Description

Function forms hierarchic clustering so that each fusion minimizes beta deiversity or gamma diversity.

Usage

diverclust(
  x,
  renyi = 1,
  equalize = TRUE,
  beta = TRUE,
  hill = FALSE,
  trace = TRUE
)

Arguments

x	Community data.
renyi	The scale of Renyi diversity index as defined in renyi. Value 0 gives diversity based on species richness, 1 diversity based on Shannon entropy, 2 diversity based on Simpson index, and Inf diversity based on Berger-Parker index. Other non-negative values are also allowed, but they may not define a well known standard diversity index.
equalize	Equalize data rows so that they can be meaningfully pooled and averaged. If this is FALSE, raw data will be used, and beta diversities may be negative. The equalization depends on the value of renyi, and each row of x is divided by (rowSums(x^renyi))^(1/renyi). In borderline cases renyi=0 the data are presence-absence tranformed (but can be left untransformed), and with renyi=Inf the rows are divided by row maxima (see decostand).
beta	Use beta diversities: the average alpha diversity of cluster members is subtracted from the pooled diversity. If this is FALSE,the clustering is based on pooled diversity, also known as gamma diversity.
hill	Use Hill numbers instead of Renyi diversity (see renyi). For renyi = 0 these are species richess values instead of their logarithms, and for renyi = 1 they are exponents of Shannon diversity instead of Shannon diversities. In general, a Hill number is an exponent of Renyi diversity. The Hill numbers may not be strictly additive, and beta diversities may be negative (except with renyi = 0).
trace	Trace calculations. Either logical or an integer: trace=2 also traces merges.

Details

The function forms clusters so that pooled diversity or its change to the baseline alpha diversity is minimized at each level of clustering. The change of pooled diversity with respect to the baseline alpha diversity is called beta diversity, and the overall diversity is called gamma diversity. For beta diversity, the clustering implies an additive partitioning.

Value

Function returns an hclust object.

Author(s)

Jari Oksanen.

See Also

hclust for cluster analysis and its support methods, renyi for estimating Renyi diversities and Hill numbers, and adipart for related additive partitioning of beta diversity.

jarioksa/natto documentation built on March 28, 2024, 12:45 a.m.