Constrained community clustering

Abstract

Constrained clustering uses pairwise constraints, i.e., pairs of data that belong to the same or different clusters, to indicate the user-desired contents. In this paper, we propose a new constrained clustering algorithm, which can utilize both must-link and cannot-link constraints. It first adaptively determines the influence range of each constrained data, and then performs clustering on the expanded range of data. The promising experiments on the real-world data sets demonstrate the effectiveness of our method. © 2014 Springer International Publishing Switzerland.