Um Método Não Supervisionado para o Povoamento de Ontologias a partir de Fontes Textuais na Web

The increasing in the production and availability of unstructured information on the Web grows daily. This abundance of unstructured information is a great challenge for acquisition of structured knowledge. Many approaches have been proposed for extracting information from texts written in natural language. However, only a few studies have investigated the extraction of information from texts written in Portuguese. Thus, this work aims to propose and evaluate an unsupervised method for ontology population using the Web as a big source of information in the context of the Portuguese language. The results of the experiments are encouraging and demonstrated that the proposed approach reached a precision rate of 67% in the instances of ontological classes extraction.