Sistema de aplicação unificada de regras linguísticas e ontologias para a extração de informações

Abstract

Information extraction is an important part of a broader set of enabling tools to assist on identifying relevant information from natural language texts. Knowledge acquisition rules, based on linguistic treatment of specific aspects of textual documents, can provide an even broader set of possibilities. This work presents a model for addressing information extraction from texts based on ontology, which uses Natural Language Processing techniques and annotated corpus to identify relevant information. The main components of the proposal are described and presented a case study based on Brazilian legal documents. The results achieved on experiments indicate relevant accuracy and precision performance and good prospects regarding flexibility, expressiveness and generalization of the extraction rules.