Research On Hypertensive Cardiovascular Risk Classifcation System Based On Bayesian Network And Ontology

Currently, the measurement of office blood pressure is the principal method to diagnose hypertension. But the method simply relied on office blood pressure may lead to misdiagnosis or missed diagnosis. Missed diagnosis, such as masked hypertension, may threat to the safety of patients. Misdiagnosis, such as "white coat effect", will cause waste of medical resources.It is by no means easy to treat hypertension, since most patients have cardiovascular risk factors along with high blood pressure. When physicians determine the treatment timing and strategy for the patients, they should not only consider blood pressure levels, but also evaluate and classify cardiovascular risk levels of the patients. But in fact, the classfication of cardiovascular risk levels of hypertensives is a complex and time-consuming work, and generally completed by experienced physicians. With the increasing number of hypertensive, the physicians have to receive more patient every day, which not only requires the physicians to complete a diagnosis in a shorter period of time, but also increases their work pressure. Since the method simply relied on office blood pressure may lead to misdiagnosis or missed diagnosis, an ontology-driven Bayesian network model is proposed in the paper in order to assist medical diagnosis. First, an ontology model used for the diagnosis of hypertension is designed. Then, relationships between ontology instances are mapped to dependencies between Bayesian network variants in order to construct a Bayesian network automatically. Finally, the Bayesian network model is used to assist the diagnosis of hypertension. Experimental results show that the model is correct and feasible.In order to assist the physicians to classify the cardiovascular risk of hypertensive, hypertensive cardiovascular risk classification system is designed in this paper to help physicians classify cardiovascular risk levels which benefits determination of treatment timing and strategy. First, electronic medical record database of the hypertension is designed to store patient medical information. Second, ontology model and knowledge base are constructed for the classification task. Third, an algorithm which can dynamically convert database records into ontology instances is proposed in order to generate ontology instances. Finally, the ontology model, knowledge base and dynamical ontology instances are banded to a reasoner to get diagnostic conclusions. Friendly graphical user interfaces are adopted in the classification system to enhance the user experience. Experimental results demonstrate that the proposed method is feasible, and can take advantage of the merits of both ontology which has strong expression ability and database which has high-efficiency access.This paper has researchful characteristics. Thus, the system proposed in the paper may have some deficiencies.