disana: Dissimilarity Analysis

In labdsv: Ordination and Multivariate Analysis for Ecology

Dissimilarity Analysis

Description

Dissimilarity analysis is a graphical analysis of the distribution of values in a dissimilarity matrix

Usage

disana(x, panel='all')

Arguments

x	an object of class ‘dist’ such as returned by dist, dsvdis. or vegdist
panel	a switch to specify which panel of graphics should be displayed. Can be either an integer from 1 to 3, or the word ‘all’.

Details

Calculates three vectors: the minimum, mean, and maximum dissimilarity for each sample in a dissimilarity matrix. By default it produces three plots: the sorted dissimilarity values, the sorted min, mean, and maximum dissimilarity for each sample, and the mean dissimilarity versus the minimum dissimilarity for each sample. Optionally, you can identify sample plots in the last panel with the mouse.

Value

Plots three graphs to the current graphical device, and returns an (invisible) list with four components:

min	the minimum dissimilarity of each sample to all others
mean	the mean dissimilarity of each sample to all others
max	the maximum dissimilarity of each sample to all others
plots	a vector of samples identified in the last panel

Note

Dissimilarity matrices are often large, and difficult to visualize directly. ‘disana’ is designed to highlight aspects of interest in these large matrices. If the first panel shows a long limb of constant maximum value, you should consider recalculating the dissimilarity with a step-across adjustment. The third panel is useful for identifying outliers, which are plots more than 0.5 dissimilar to their nearest neighbor.

Author(s)

David W. Roberts [e-mail]

Examples

data(bryceveg) # returns a data.frame called veg
dis.bc <- dsvdis(bryceveg,'bray/curtis')
disana(dis.bc)

labdsv documentation built on April 10, 2023, 5:08 p.m.