1. GENERALIZATIONS OF K-MEANS ALGORITHMS FOR CONSTRAINED CLUSTERING
Authors: HARUKA KONDOH , SADAAKI MIYAMOTO
Abstract: A generalization of K-means clustering algorithms including cluster size variables and covariance variables are introduced. Moreover an on-line version of that is considered. A constrained clustering algorithm using these generalizations and the idea in the COP K-means is proposed. Performances of the proposed algorithms are compared using numerical examples.
Keywords: Generalized K-means clustering, on-line algorithm, constrained clustering