bjdbrda: Multiple dbRDA from Beta Distributed Jaccard Dissimilarity

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

Multiple dbRDA from Beta Distributed Jaccard Dissimilarity

Description

Function performs distance-based RDA on expected Beta Jaccard dissimilarities, and then reruns the analysis on given number of random Beta Jaccard dissimilarities.

Usage

bjdbrda(formula, data, n = 100, ...)

## S3 method for class 'bjdbrda'
scores(
  x,
  choices = 1:2,
  display = c("wa", "lc", "bp", "cn"),
  scaling = "species",
  const,
  expected = TRUE,
  ...
)

## S3 method for class 'bjdbrda'
plot(
  x,
  choices = 1:2,
  wa = "p",
  lc = "n",
  cn = "hull",
  bp = "wedge",
  wa.par = list(),
  lc.par = list(),
  cn.par = list(),
  bp.par = list(),
  scaling = "species",
  type = "t",
  ...
)

## S3 method for class 'bjdbrda'
boxplot(
  x,
  kind = c("eigen", "correlation"),
  points = "red",
  pch = 16,
  xlab = "Axis",
  ylab,
  ...
)

Arguments

formula, data	Model definition of type Y ~ Var1 + Var2, data = X, where Y is dependent community data (handled as binary data), Var1 and Var2 are independent (explanatory) variables found in data frame X. See dbrda for further information.
n	Number of Jaccard dissimilarity matrices sampled from Beta distribution.
...	Other parameters passed to functions; passed dbrda in bjdbrda, ignored in scores.
x	bjdbrda result object.
choices	Selected ordination axes.
display, scaling, const	Kind of scores, the scaling of scores (axes), and scaling constant with similar definitions as in scores.rda.
expected	Return scores of the expected ordination instead of ordination based on random samples from Jaccard dissimilarity.
wa, lc, cn, bp	Display of corresponding scores. "n" skips the score, "p" and "t" use points or text for the expected score, and other shapes define bjpolygon or bjstars shape used for sampled random scores.
wa.par, lc.par, cn.par, bp.par	List of arguments to modify the plotting parameters of the corresponding shape.
type	Add "t"ext or "p"oint for the expected score to a shape of scatter of random shapes.
kind	Draw boxplots of eigenvalues or pairwise axis correlations for each random sample.
points, pch	Add points of given color and shape to the eigenvalue boxplot.
xlab, ylab	Change labelling of axes in boxplot.

Details

Function works only with constrained and partial constrained ordination, and only saves the randomized results of constrained ordination. To maintain the eigenvalues, function does not rotate the random results to the reference ordination, but it will reflect axes with reversed signs. The eigenvalues and magnitudes of axiswise correlations to the reference ordination can be inspected with boxplot.

The display type in plot can be specified independently to each type of score (or the score can be skipped). The default display can be modified using a similarly named list of graphical argument values that will replace the defaults. For instance, to display linear combination scores as convex hull, use lc = "hull", and modify its parameters with list lc.par. For available graphical parameters, see bjpolygon for filled shapes, bjstars for stars, points for points, and ordilabel for text. Alternatives "p" and "t" will only show the scores of the reference solution.

Value

Function returns dbrda result object amended with item BayesJaccard that is a list of randomized results with following items for each random sample:

• tot.chi total inertia

• eig eigenvalues of constrained axes

• r absolute value of correlation coefficient of randomized axis and the reference axis

• u, wa, biplot, centroids ordination scores for linear combination scores, weighted averages scores, biplot scores and centroid of factor constraints; these are unscaled and are usually accessed with scores.bjdbrda for appropriate scaling

Examples

if (require(vegan)) {
data(dune, dune.env)
m <- bjdbrda(dune ~ A1 + Management + Moisture, dune.env)
## eigenvalues and correlations showing axis stability
boxplot(m)
boxplot(m, kind = "correlation")
## Default plot
plot(m)
## modify plot
plot(m, cn = "star", wa = "ellipse", cn.par = list(col=gl(2,4)),
   wa.par=list(col = dune.env$Management, keep = 0.607))
}

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