Learning Pairwise Similarity for Data Clustering

Each clustering algorithm induces a similarity between given data points, according to the underlying clustering criteria. Given the large number of available clustering techniques, one is faced with the following questions: (a) Which measure of similarity should be used in a given clustering problem? (b) Should the same similarity measure be used throughout the d-dimensional feature space? In other words, are the underlying clusters in given data of similar shape? Our goal is to learn the pairwise similarity between points in order to facilitate a proper partitioning of the data without the a priori knowledge of k, the number of clusters, and of the shape of these clusters. We explore a clustering ensemble approach combined with cluster stability criteria to selectively learn the similarity from a collection of different clustering algorithms with various parameter configurations