IEEE Computer Society Newsfeed

Subscribe RSS

« Back

New Algorithm Detects Virtual Communities

University of Ljubljana researchers have created a new algorithm they say can better detect virtual communities than conventional approaches. The researchers define communities as systems of nodes interacting through links. This might include social groups or even protein-protein interactions in yeast. The researchers developed a propagation-based algorithm that recognizes communities within large amounts of network data based on the link density and link patterns. It is able to find these communities within network data without any prior knowledge of the number of communities involved. They tested the algorithm on 10 real networks—including social and biological networks—and found it detected communities more accurately than typical algorithms. Applications for the new algorithm could include predicting future friendships in online social networks, analysis of interactions in biological systems that are typically difficult to observe, and detecting duplicate software code. The researchers published their work in European Physical Journal B. (EurekAlert)(Springer Select)(L. Šubelj and M. Bajec, “Ubiquitousness of link-density and link-pattern communities in real-world networks,” European Physical Journal B (EPJ B) 85: 32, (2012).)

Trackback URL: