Classic: A hierarchical clustering algorithm based on asymmetric similarities

摘要

The nearest neighbors relation (NNR) is defined in terms of a given asymmetric matrix of similarities of data items. This paper presents a new clustering algorithm, called CLASSIC, based on an iteratively defined nested sequence of NNRs. CLASSIC has been applied to various types of gestalt clustering problems. For CLASSIC applications in which asymmetric similarities are not available a priori, this paper also introduces a method for obtaining asymmetric similarities from Euclidean distances. This method has been used in the detection of gestalt clusters by CLASSIC.