PCPS (version 1.0.7)

pcoa.sig: Significant dimensions in principal coordinate analysis

Description

Function for determine the number of significant dimensions in principal coordinate analysis (PCoA).

Usage

pcoa.sig(
  data,
  method = "gower",
  squareroot = FALSE,
  axis = 6,
  n.start = NULL,
  by = 1,
  iterations = 1000,
  parallel = NULL
)
# S3 method for pcoasig
print(x, ...)
# S3 method for summarypcoasig
print(x, ...)
# S3 method for pcoasig
summary(object, choices = c(1, 2), ...)

Arguments

data

Community data matrix.

method

Method for dissimilarity index, as accepted by vegdist (Default method = "gower").

squareroot

Logical argument (TRUE or FALSE) to specify if use square root of dissimilarity index (Default squareroot = FALSE).

axis

Maximum number of ordination principal axes to be monitored (Default axis = 6).

n.start

Initial sample size. If n.start = NULL initial sample size is equal to total sample size (Default n.start = NULL).

by

Sampling unit is added at each sampling step (Default by = 1).

iterations

Number of permutations to assess significance (Default iterations = 1000).

parallel

Number of parallel processes or a predefined socket cluster done with parallel package. Tip: use detectCores() (Default parallel = NULL).

x

An object of class pcoasig.

...

Other parameters for the respective functions.

object

An object of class pcoasig.

choices

Axes for re-scaling. Choices must have length equal to two (Default choices = c(1, 2)).

Value

value

The eigenvalues, relative eigenvalues and cumulative relative eigenvalues..

vectors

The principal coordinates.

correlations

Correlations between axis and original data.

mean.cor.null

Mean correlations, for axis, between null and reference scores.

mean.cor.bootstrap

Mean correlations, for axis, between bootstrap and reference scores.

n.permut.bootstrap

Number of iterations for each axis in bootstrap step.

n.permut.null

Number of iterations for each axis in null step.

probabilities

Probabilities for each axis.

Details

At each iteration step a bootstrap sample is subjected to PCoA ordination, the scores are submitted to a procrustean adjustment, and the correlation between observed and bootstrap ordination scores is computed. It compares such correlations to the same parameter generated in a parallel bootstrapped ordination of randomly permuted data. The number of axes in bootstrap or null PCoA with eigenvectors corresponding to positive eigenvalues may be smaller than the number of axes monitored, in this case, axes with values equal to 0 are created. The number of iterations with original values for each axis is shown in n.permut.bootstrap and n.permut.null.

The function scores.pcoasig re-scales the correlation values for biplot graphics.

References

Pillar, V.D. (1999). The bootstrapped ordination reexamined. Journal of Vegetation Science 10, 895-902.

See Also

pcoa, procrustes

Examples

# NOT RUN {
# }
# NOT RUN {
data(flona)
res<-pcoa.sig(flona$community, method = "bray", squareroot = TRUE, axis = 6, iterations = 100)
res
summary(res)$scores
# }
# NOT RUN {
# }