Automatic Construction of Taxonomies of Categories

Hierarchies are an intuitive and effective organization paradigm for data. Of late there has been considerable research on automatically learning a hierarchical organizations of data. In this paper, we formulate the problem of “automatically constructing hierarchical taxonomies”, which we define as learning a hierarchy of categories with no user defined parameters. We propose a framework that characterizes a “good” taxonomy and also provide an algorithm to find it. Finally, we evaluate our approach on multiple real life datasets from diverse domains, such as text mining, hyper-spectral analysis, written character recognition etc. We show that our approach constructs “better” taxonomies than other approaches, and that the output space decompositions induced by these taxonomies yield better classification accuracies.