A Pairwise Distance View of Cluster Validity

Ensemble Clustering based on Heterogeneous Dimensionality Reduction Methods and Context-dependent Similarity Measures

“…Although a common approach used in pattern recognition, clustering remains a somewhat ill-defined problem: the objective is to find good groupings/clusters of data point with little or no information on the ground truth. A discussion of the issues underlying cluster validity can be found in [17]. The set of data vectors that satisfy certain criteria of proximity are grouped into clusters to reduce information load and facilitate information processing and interpretation.…”