[R] Grouping on a Distance Matrix

Dario Strbenac dstr7320 at uni.sydney.edu.au
Thu Feb 13 09:00:12 CET 2014


I'm looking for a function that groups elements below a certain distance threshold, based on a distance matrix. In other words, I'd like to group samples without using a standard clustering algorithm on the distance matrix. For example, let the distance matrix be :

      A     B     C     D
A     0  0.03  0.77  1.12  
B  0.03     0  1.59  1.11
C  0.77  1.59     0  0.09  
D  1.12  1.11  0.09     0

Two clusters would be found with a cutoff of 0.1. The first contains A,B. The second has C,D. Is there an efficient function that does this ? I can think of how to do this recursively, but am hoping it's already been considered.

Dario Strbenac
PhD Student
University of Sydney
Camperdown NSW 2050

More information about the R-help mailing list