Discovering Social Groups without Having Relational Data

摘要

Who is associated with whom? Who communicates with whom? When two or more individuals get together is there an intended purpose? Who are the leaders/important individuals of the group? What is the organizational structure of the group? These are just a few of the questions that are covered under the topic of social network analysis. Data mining, specifically community generation, attempts to automatically discover and learn these social models. In this paper we present one class of problems which we have called the uni-party data community generation paradigm. We discuss various applications, a methodology and results from two problem domains.