Finding scientific communities in citation graphs: Articles and authors

Understanding the nature and organization of scientific communities is broad interest. The `Invisible College' a historical metaphor for one such type community search `colleges' can be framed as detection analysis small groups scientists working on problems common interests. Case studies have previously been conducted individual with respect to their social behavior. In this study, we introduce, new scalable finding approach. Supplemented by expert assessment, use convergence two different clustering methods select article clusters generated from over million articles field immunology spanning an eleven year period relevant cluster quality indicators evaluation. Finally, identify author defined these clusters. A sample produced pipeline was reviewed experts, shows strong thematic relatedness, suggesting that inferred may represent valid practice. These findings suggest convergent approaches useful in future.