Algoritmo flexibilizado de agrupamiento semántico

Abstract

In this paper, we propose a flexible algorithm for semantic clustering. We introduce equal couple, semi-equal couple, null couple, semi-null couple, match couple, binding and semantic couple. The algorithm matches the words of two strings by a set of operations (insertion, deletion and substitution). We determine the minimum cost for each operation required to change one string into another, using the alignment algorithm of Wagner & Fisher. Following these transformations, we align pairs of words for two definitions, to obtain bindings (then strongest pairs of words candidates for semantic couples), and the semantic clusters (semantically related words in a given context). The flexible algorithm for semantic clustering was implemented into a system accessible on-line at: http://iling.torreingenieria.unam.mx.