Dear SOCNET members,

In social network analysis, a signed graph is clusterable if it can be partitioned into three or more clusters such that all positive ties are within the clusters and all negative ties are situated between clusters. Note, this is in contrast to balance, whereby a signed graph is balanced if it can be partitioned into two clusters (one of which may be empty) such that all positive ties are within the clusters and all negative ties are situated between clusters. In other words, clusterability is the more general model in terms of which balance is a special case

I would be grateful if anyone can inform me about the first appearance of a clusterability procedure or algorithm in the network (or graph theory) literature.

Thank you,

Ion Georgiou