Uma avaliação do raciocínio baseado em casos aplicado a recuperação de jurisprudências no Tribunal Regional Eleitoral do Distrito Federal

AUTOR(ES)
DATA DE PUBLICAÇÃO

2009

RESUMO

This is a research whose object is to evaluate a legal information retrieval system precision. This IR system is based on a model that uses artificial intelligence technique known as Case- Based Reasoning (CBR). In this model the jurisprudences are organized in the form of actual legal cases. The principle of CBR is that a past legal case can be useful to solve a current problem, since there is between them some degree of similarity. To establish similarities between current and past cases the model studied proposes the use of the similarity calculation performed based on comparison of thematic indices. The process of indexing is performed by experts using a thesaurus as a legal support. This research uses jurisprudences produced by the Regional Electoral Tribunal of the Distrito Federal. The sample was composed considering electoral jurisprudence produced in general elections of Distrito Federal, in the year of 2006. To perform the evaluation of the model, a prototype of a casebased information retrieval system was built. Then the prototype precision degree was evaluated from the result of a set of queries submitted to it. The method adopted for the evaluation was the same used in the Text REtrieval Conference (TREC) in 2007 by Legal Track Task. After the data collecting, a report was made to discuss the possibility of the casebased information retrieval system can be considered a paradigm for the legal information retrieval.

ASSUNTO(S)

recuperação da informação ciencia da informacao similarity calculation information retrieval avaliação de sistemas de recuperação da informação case-based reasoning information retrieval system evaluation electoral jurisprudence cálculo de similaridade raciocínio baseado em casos jurisprudência eleitoral

ACESSO AO ARTIGO

Documentos Relacionados