Hello,
I was interested in trying to use community similarity (rather than environmental similarity) for creating the site weights for the Frescalo analysis, as Hill (2012) does. I implemented it locally quite simply in the createWeights() function, first adding an argument to the function choosing the similarity method (with Euclidean as default to preserve current functionality):
I also added a relevant check to errorChecks() regarding choice of distance function, and I guess vegan would need to be added to the dependent packages.
Is this something useful enough that you might be interested in implementing into the package? If I can manage to work out how to do pull requests properly then I'm happy to try it, but I'm also open to you maintainers doing it yourselves.
Hello, I was interested in trying to use community similarity (rather than environmental similarity) for creating the site weights for the Frescalo analysis, as Hill (2012) does. I implemented it locally quite simply in the
createWeights()
function, first adding an argument to the function choosing the similarity method (with Euclidean as default to preserve current functionality):createWeights<-function (distances, attributes, sim_method="euclidean", dist_sub = 200, sim_sub = 100, normalise = FALSE)
Then replacing the current method using the
dist()
function withvegan
'svegdist()
:sim_distance <- vegdist(attributes[, 2:length(names(attributes))], method=sim_method, diag=TRUE, upper=TRUE)
I also added a relevant check to
errorChecks()
regarding choice of distance function, and I guess vegan would need to be added to the dependent packages.Is this something useful enough that you might be interested in implementing into the package? If I can manage to work out how to do pull requests properly then I'm happy to try it, but I'm also open to you maintainers doing it yourselves.
Thanks, Alistair