Técnicas de recuperación de información aplicadas a la construcción de tesauros

Abstract

The aim of the article was to propose the application of a set of techniques used in Information Retrieval for the development of a Thesaurus. The proposed ideas have been applied in the selection of the terminology; categorization of terms by creating clusters; and establishment of semantic relationships between terms through semantic similarity, which resulted in a Foreign Trade Thesaurus of 7,790 terms. From these results, we concluded that the techniques used significantly simplified the tasks of obtaining the terminology, and they can improve the quality of the final thesaurus. In addition, the techniques enabled the analysis of the conditions of the collection for which the thesaurus is used and provide extra information that would be hard to obtain manually.