| pcdists {pcurve} | R Documentation |
Calculates and returns a matrix of extended dissimilarities between rows of a data matrix using the specified distance measure.
pcdists(data, dcrit = 1, use.min = TRUE, metric = "man", scale = TRUE, eps = 0.0001, replace.neg = TRUE, big = 10000)
data |
numeric data matrix |
dcrit |
critical distance, the threshold dissimilarity at which the dissimilarity is replaced using the flexible shortest path adjustment method. |
use.min |
|
metric |
distance measure to be calculated. "man",
Manhattan (= row-standardized Bray-Curtis); "euc", Euclidean;
"bin", Binary. see dist for further details. |
scale |
if TRUE each column is standardized to the maximum value, if FALSE, columns remain unstandardized. |
eps |
minimum difference from critical distance for iterations. |
replace.neg |
|
big |
This function calculates calculates extended dissimilarity matrices, including the Bray-Curtis distance measure, of common application in ecology. See De'ath (1999) for a full discussion of extended dissimilarity.
A list comprising
d |
a dissimilarity matrix of the specified metric. |
dnew |
a matrix of extended dissimilarity measures. |
R port by Chris Walsh Chris.Walsh@sci.monash.edu.au from S+ library by Glenn De'ath glenn.death@jcu.edu.au.
De'ath, G. 1999 Extended dissimilarity: method of robust estimation of ecological distances with high beta diversity. Plant Ecology 144, 191–199.
#for simulated data set with high Beta diversity (many rows with #no columns in common, and then calculate extended dissimilarity #matrices with critical distance = 1 and 0.7. Finally plot #dissimilarities in each case against ecological distance to #demonstrate the effect of extended dissimilarities on the #linearity of the relationship. data(sim10var) species <- sqrt(sim10var[, 2:11]) specdist <- pcdists(species, dcrit = 1, metric = "man", scale = FALSE) SD <- c(specdist$d) XD <- c(specdist$dnew) specdist <- pcdists(species, dcrit = 0.7, metric = "man", scale = FALSE) XD1 <- c(specdist$dnew) #Ecological distance as euclidean distance between locations #on known gradient ecdist <- dist(sim10var[, 1], method = "euc") ED <- c(ecdist) par(mfrow = c(1, 3)) plot(ED, SD) plot(ED, XD) plot(ED, XD1) par(mfrow = c(1, 1))