Comparing large datasets structures through unsupervised learning

Abstract

In data mining, the problem of measuring similarities between different subsets is an important issue which has been little investigated up to now. In this paper, a novel method is proposed based on unsupervised learning. Different subsets of a dataset are characterized by means of a model which implicitly corresponds to a set of prototypes, each one capturing a different modality of the data. Then, structural differences between two subsets are reflected in the corresponding model. Differences between models are detected using a similarity measure based on data density. Experiments over synthetic and real datasets illustrate the effectiveness, efficiency, and insights provided by our approach. © 2009 Springer-Verlag Berlin Heidelberg.