Sistema de recomendación para textos legales

Abstract

En el presente trabajo realizamos una solución para el desarrollo de un sistema de recomendación de documentos de texto el cual se lo instancia al dominio legal/jurídico, utilizando el corpus de leyes de Argentina, accesible desde la página web oficial www.infoleg.gob.ar. Realizamos la investigación y desarrollo de algunos motores para recomendación de texto, junto a una plataforma para visualizar recomendaciones. Posteriormente se analizan cualitativamente los resultados obtenidos en cada caso.

Recommender systems have been an important and even essential component in several platforms. Its main attractive is that it gives relevant information to the user in an active way, about the database in question, without the need for it to have knowledge about the recommended articles or the query to be made. The value of this information comes from the previous analysis of the data, and its possible relationship with users. In the present work we made a solution for the development of a recommendation system of text documents which we instantiate to the legal/juridical domain, using the corpus of laws of Argentina, accessible from the official website www.infoleg.gob.ar. We carry out research and development of some engines for text recommendation, along with a platform to view the recommendations. Subsequently, the results are analyzed qualitatively in each case.