A indexação temática de recursos fundamentada por estrutura profunda eabordagem objeto-relacionamento

Abstract

This study was motivated by the problem of electronic resources recovering without the semantic hierarchy describing their conceptual field. The indexer activity was seen as the absence of semantic representation structure construction with some potential to be processed by machines and people. The bibliographical revision has been conducted from the assumption that there would be ideal semantic structures to represent those resources, both in the area of Information Science and in the area of Computer Science, considering two different theoretical approaches. The first, coming from the general theory of subject classification, proposes the deep structure DEPAm (Discipline, Entity, Property, Action, modifier). The second, coming from the object-oriented modeling, proposes the ER (Entity-Relationship) modeling, which involves a triple representation and a description based on scheme and syntax RDF/XML (Resource Description Framework/Extended Markup Language). Upon such theories, the subject indexing model DEPAm-OR was proposed. Through the methodology involving models and content analysis, it was made the data collection and the analysis of the electronic resources made available by information services. The implementation of the ObjectRelationship-OR structure was became possible though the free softwares, HiperEeditor and HiperNavegador, employment. It has been concluded that the proposed DEPAm-OR model is an important tool. Its usage is recommended since it allows a simple representation of the resource subjects. lt also adds them a great semantic explicitnesssuch as it is determined by the deep structure conventions.