Bray-Curtis for more than pairwise overlap
Jari, Thanks, I'll look at those. To answer your questions below 1) I'm looking at negative values as I'm not comparing dissimilarity in composition, but dissimilarity in species effects. Sometimes they are negative. I suppose I could add a constant to bring the lowest value to 0, but that seems incorrect. 2) I have actually written new R code, so, I'm sure its slower, but, I needed to do so in order to handle these cases. I'll take a look at the methods you implemented to see if they'll work and/or compare to the multi-case derivation I've come up with. -Jarrett
On May 4, 2012, at 10:15 PM, Jari Oksanen wrote:
Jarrett, I really do not understand what you try to do and what properties you expect your measure to have. However, Baselga (Global Ecol Biogeog 19, 134-143; 2010) has proposed multiple site S?rensen (= Bray-Curtis for presence/absence) and Jaccard indices, and they are implemented as functions nestedbetasor() and nestedbetajac() in vegan (recent version).
Great, I'll look at these.
On 05/05/2012, at 00:48 AM, Jarrett Byrnes wrote:
As implemented in the literature, Bray-Curtis dissimilarity measures are used to look at pairwise comparisons. So, how dissimilar are 2 plots in their composition, for example. I'm wondering if anyone has run across or implemented a method to look at BC dissimilarity or otherwise that has two different twists to it 1) Negative values are possible.
What do these negative values mean? Do they mean that compared groups are more similar than absolutely identical? Methods that expect to have dissimilarities as input will not usually accept negative dissimilarities (but these can appear in results when analysing non-Euclidean indices, but that's another issue).
2) The index is for an n-wise measure of dissimilarity. For example, instead of asking how dissimilar 2 plots are, what is the dissimilarity between 3 plots. I've been trying to play with vegdist to see if I can implement this, and have been rolling my own functions that work well with just presence-abscence data. But I'm unclear of how to do so for continuous data. Thanks!
In vegdist, the actual BC is calculated in C code. Do you mean that you have written new C code for this special variant or are you deriving your index from pairwise dissimilarities? Cheers, Jari Oksanen -- Jari Oksanen, Dept Biology, Univ Oulu, 90014 Finland