This Collaborative REU project, which is being conducted at Pomona College, is sponsored by the Computing Research Association Committee on the Status of Women in Computing Research (CRA-W) and the Coalition to Diversify Computing (CDC). During the 2015-2016 academic year, this project is funded by the National Science Foundation. During Summer 2015, this project received funding from Pomona College’s Summer Undergraduate Research Program (SURP). More information about the participants can be found here.
Complex networks can be used to represent interactions between members of a set. A community within a network is a set of members that are more connected to each other than to other members. In many applications, we are interested in identifying communities within these networks based solely on the interactions observed. The majority of commonly-used algorithms for detecting communities optimizes a metric known as modularity, which compares the density of connections of members within a community to members of other communities. However, other quality metrics such as conductance, coverage, performance, and silhouette index could also be used within those same algorithms. For this study, we will examine the impact of replacing modularity with these quality metrics in existing implementations of the Louvain and CNM community detection algorithms.