Authors:
Lorena S. C. de Oliveira
;
Rodrigo V. Andreão
and
Mário Sarcinelli-Filho
Affiliation:
Federal University of Espírito Santo, Graduate Program on Electrical Engineering, Brazil
Keyword(s):
Artificial Intelligence, Medical Informatics, Bayesian Networks, Decision-Support Systems, PVC detection.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Biomedical Engineering
;
Decision Support Systems
;
Expert Systems
;
Health Information Systems
;
Knowledge Engineering and Ontology Development
;
Knowledge-Based Systems
;
Symbolic Systems
Abstract:
This paper presents a system based on Bayesian networks (BN) to support medical decision-making. The proposed approach is able to learn from available data, and provides an intuitive graphical interpretation of the problem, which can be easily configured by a physician. This approach is evaluated for the first time in the problem of premature ventricular contraction (PVC) detection, using a representative set of records of the MIT-BIH database. The results obtained emphasize the capability of the Bayesian network to make decisions even when the information about some symptoms or events is not complete. Moreover, the good performance obtained opens many perspectives for the use of BN to deal with beat classification.