Recuperação de documentos baseados em informação semântica no ambiente AMMO.

Abstract

This study presents techniques used for extracting semantic information from texts and strategies for semantic information-based document retrieval. These techniques and strategies have been adopted so as to develop a Document Manipulation System based on Semantic Information. Previously established values of similarity and relevance are used in the document retrieval process.. The organization of semantic information, user interfaces, manipulation of imprecise information and the extraction and retrieval mechanism are discussed. The impreciseness in the semantic information is treated through similarity values in order to compare the query´spredicates with the database results. Relevance values specified for the query´s predicates have been used to classify the resulting documents. A prototype of the system has been developed. This prototype can be incorporated into the AMMO environment (Authoring and Manipulation of Multimedia Objects), whose aim is to provide resources for creating, storing and manipulating multimedia applications.