EcotoneFinder (version 0.2.3)

ExtractCentroid: Visualisation of fuzzy centroids:

Description

Visualisation of fuzzy centroids:

Usage

ExtractCentroid(ecotonefinder, method = c("fanny", "cmeans", "vegclust"),
  normalized = c("species", "cluster", "none"), position = "dodge",
  threshold = 0, plot = TRUE, col = NULL, return.plot = TRUE,
  labels = ggplot2::waiver(), main = "Community composition",
  xlab = "species", ylab = "Centroid contribution", cex.x = 12)

Arguments

ecotonefinder

A list containing elements named in the same way than EcotoneFinder function outcomes. Must contain <U+201C>cmeans<U+201D>, <U+201C>fanny<U+201D> or <U+201C>vegclust<U+201D> results.

method

The fuzzy clustering results from which the centroids will be extracted.

normalized

Method to normalise the centroid values, either by <U+201C>species<U+201D> or <U+201C>cluster<U+201D>. If <U+201C>none<U+201D>, the centroids are plotted without transformation. See details.

position

Set the positions of the bars for the barchart. This is passed down to the geom_bar function of ggplot. Default is set to <U+201C>dodge<U+201D>.

threshold

Threshold for centroid contribution value under which the species will not be plotted. Can be used to simplify plots containing many species. See Details.

plot

Logical. Should the plot be displayed. If FALSE, the centroids matrix is returned without plotting.

col

Colour vector for the plot. Should be of the same length that the number of fuzzy clusters.

return.plot

Logical. Should the GGplot object be stored internally (e.g. for multi-ploting). Default is TRUE.

labels

Character vectors of labels for the legend. Must be of the same length that the number of fuzzy clusters.

main

Main title for the plot. See plot.

xlab

A title for the x-axis. See plot.

ylab

A title for the y-axis. See plot.

cex.x

cex for the x-axis labels.

Value

A matrix containing the cluster centroids.

Details

This function extracts and plots the fuzzy centroids species contributions, according to user-defined normalisation steps and threshold value. The contributions of the different species in the fuzzy centroids may be used as a proxy for community compositions. The cmeans function (cmeans package) and vegclust function (vegclust package) internally compute the centroid compositions and their outputs are directly used by the ExtractCentroid function. The fanny function (cluster package), however, does not provide internal centroids calculation. They are computed here as: $$Centroid[cluster j] = \sum[ij] (Membership[ij] x Observation[ij]) / \sum[j] Membership[j] $$ Where the centroid of a cluster is the mean of all observations, weighted by their degree of belonging to the cluster. The obtained species contributions to the centroids of the fuzzy clusters can then be plotted as they are, if normalised = <U+201C>none<U+201D>. To obtain more intuitive units for the interpretation of the species contributions, two normalisation methods are proposed. If normalised = <U+201C>cluster<U+201D>, the species contributions are given in percent per clusters (i.e. the sum of all species contributions in each cluster centroid equals 100). If normalised = <U+201C>species<U+201D>, each species has its contributions summed to 100 (i.e. each species is in percent per cluster). For normalised = <U+201C>none<U+201D> and normalised = <U+201C>cluster<U+201D>, a threshold value can be specified. Species that do not score above this threshold will not be displayed on the resulting plot. This can be used to simplify the outputs, for dataset containing large number of species.

Examples

# NOT RUN {
##### Artificial dataset & analyses:
 SyntheticTrial <- SyntheticData(SpeciesNum = 20, CommunityNum = 4,
                                 SpCo = NULL ,Length = 500,
                                 Parameters=list(a = rep(60, 4),
                                                 b = c(0,150,350,500),
                                                 c = rep(0.015,4)),
                                 dev.c = 0.007,
                                 pal = c("#008585", "#B8CDAE", "#E6C186", "#C7522B"))

 EcoFinder <- EcotoneFinder(SyntheticTrial[,-1],
                            dist = SyntheticTrial$Distance,
                            method = "all", groups=4,
                            standardize = "hellinger",
                            diversity="all")

 ##### Centroid plot without normalisation:
 Centroid <- ExtractCentroid(EcoFinder, method = "fanny",
                             normalized = "none",  threshold = 0,
                             plot = TRUE, position = "dodge",
                             col = colorspace::heat_hcl(4))

 ##### Centroid plot normalised by clusters:
 Centroid <- ExtractCentroid(EcoFinder, method = "fanny",
                             normalized = "cluster",  threshold = 0,
                             plot = TRUE, position = "dodge",
                             col = colorspace::heat_hcl(4))

# }