Recuperação de Documentos Jurídicos Baseada em um Tesauro

The methods of information retrieval in textual bases, broadly used by the search machines, are based on techniques returned for collections of generic documents. In specific domains, as the juridical, the direct application of these methods takes the results of smaller quality than the expected. The basic reason is that such methods don't consider semantic information associated to the domain in subject. Particularly, in the case of the Brazilian Right, discussed in this article, semantic information can be obtained of the thesaurus elaborated by the Conselho de Justiça Federal (CJF) and of the structure of the juridical document. In this work, we explored the use of this thesaurus in a juridical document retrieval model. Through experimentation, we showed that this model takes the best precision (quality of the answers) that models returned for generic collections.