compart {bipartite} | R Documentation |
Finds number of compartments, based on multivariate ordination techniques, and labels interactions according to the compartment they belong to.
compart(web)
web |
A bipartite interaction web, i.e.~a matrix with higher (cols) and lower (rows) trophic levels. |
Internal function, to be called by networklevel
.
The matrix web
is “diagonalised” through CA-scores, then compartments are
detected based on the distconnect
-function from vegan, and the interactions
are assigned to the compartments. Finally, the original data structure is reinstated and returned.
The logic of using a CA before the analysis was based on a sentence somewhere in Lewinsohn et al.~(2006).
Returns a list with two entries:
cweb |
A matrix similar to web , but now with compartment numbers instead of interaction values. |
ncompart |
The number of compartments. |
Carsten F. Dormann <carsten.dormann@ufz.de>
Lewinsohn, T. M., P. I. Prado, P. Jordano, J. Bascompte, and J. M. Olesen. 2006. Structure in plant-animal interaction assemblages. Oikos 113, 174–184
See also networklevel
, distconnected
and cca
.
# make a nicely compartet web: web <- matrix(0, 10,10) web[1,1:3] <- 1 web[2,4:5] <- 1 web[3:7, 6:8] <- 1 web[8:10, 9:10] <- 1 web <- web[-c(4:5),] #oh, and make it asymmetric! web <- web[,c(1:5, 9,10, 6:8)] #oh, and make it non-diagonal compart(web) # or, standard, use Safariland as example: data(Safariland) compart(Safariland)