Generalizing the kk-Windows clustering algorithm in metric spaces

Clustering methods are one of the key steps that lead to the transformation of data to knowledge. Clustering algorithms aim at partitioning an initial set of objects into disjoint groups (clusters) such that that objects in the same subset are more similar to each other than objects in different groups. In this paper we present a generalization of the kk-Windows clustering algorithm in metric spaces. The original algorithm was designed to work on data with numerical values. The proposed generalization does not assume anything about the nature of the data per se, but only considers the definition of a distance function over the dataset. The efficiency of the proposed approach is demonstrated in various datasets.