JUSTIÇA - Automatic acquisition of facts that occurred and are described by the decisions of the Supreme Court and Natural Language Interrogation about those facts

The main result of this project was the definition of an architecture for an information retrieval system with a cooperative dialogue manager using domain knowledge represented in levels of knowledge. The knowledge levels are obtain from different sources: rules modeling juridical knowledge; rules and facts resulting from the automatic processing (syntactic, semantic and pragmatic analysis) of the descriptive parts of some text documents; facts and rules obtain by automatic building documents clusters with significant labels associated; and the information retrieval level that include the words and expressions that are in each document.

Funding entity: Fundação Ciência e Tecnologia (MCTES).