Metric of Hierarchical Choices and Possibilities of its Application in Computational Linguistics

Abstract

We are defining a new distance -the distance of hierarchical choices, and proofing that it's a metric. It is defined on a space of results of hierarchical clustering, i.e., binary trees with labelled leafs and unlabelled inner nodes, socalled dendrograms. This new metric can be used in many domains of computational linguistics, for example, for expert evaluation of idiomometric results.