Mapeamento ontológico com aplicação no domínio biomédico

Abstract

Introduction: The health area produces daily a large volume of data that must be stored efficiently. In order for the information produced to be made available, it must be organized. In this sense, technologies focused on data collection, storage and manipulation have been evolving through automated computational techniques, methods and tools. One of the most used forms is known as ontology, which allows the representation of a set of concepts and portrays the semantics of information. However, the full use of ontologies in computational systems is still restricted. Considering this constraint and based on tha fact that relational databases provide several benefits such as scalability for queries, robustness, performance, maturity, availability and reliability; an alternative that have arisen is the ontological mapping. That is, developing mechanisms in the relational database that perform functions close to ontologies and still safeguard the data. In this context, it was observed the need to develop an approach to map an ontology from the biomedical domain to a relational database so that it would assist the decision making process in the diagnosis of chronic kidney diseases (CKD). Objectives: The general objective of this proposal is to present an approach for ontology mapping in the biomedical domain for relational databases with clinical decision support and emphasis on the diagnosis process of CKD. Methods: In order to develop this study, methodological stages were defined, involving knowledge construction, data modeling, mapping execution and validation of the developed technique. Initially, the bibliographical survey was carried out to refine the necessary knowledge and to understand the object under study. Then the mapping rules were elaborated and the data modeling stages were executed, resulting in the ontological mapping. Results: The main contribution presented is DB-Ontology, a relational database to support clinical decision in the diagnosis process of CKD. This approach allows the persistence of data in the database and preserves the semantics of the biomedical domain ontology. DB-Ontology is a result of the proposed ontological mapping. Conclusion: With the implementation of the steps defined in the methodology, it was possible to map the main classes of OntoDecideDRC to DB-Ontology. In addition, it was possible to understand the reality in USFs, so that such reality would be adapted to the hierarchy of the ontology and consequently reflect in the relational database that was developed. Although the literature presentes lacks regarding ontological mapping studies using stored procedures, it was possible to develop an efficient and appropriate approach to support clinical decision making.