CAJ | 학술논문

目的：提出基于领域粗糙集的贝叶斯网络医学数据挖掘模型，探讨肝炎肝硬化的临床分类。方法：根据所收集的355例肝硬化患者临床资料，采用领域粗糙集算法提取与肝炎肝硬化临床分类有关的生物检测指标。然后，运用树增强型贝叶斯分类器构建分类模型进行肝炎肝硬化的临床分类。结果：采用领域粗糙集贝叶斯网络分类模型进行肝炎肝硬化代偿性分类的正确率为90．91％，活动性分类正确率为94．09％，而使用BP神经网络的代偿性分类正确率为76．82％，活动性分类为85．45％。结论：领域粗糙集贝叶斯网络分类方法可以有效地进行肝炎肝硬化临床分类，并能够为临床医学诊断研究提供参考。
목적：제출기우영역조조집적패협사망락의학수거알굴모형，탐토간염간경화적림상분류。방법：근거소수집적355례간경화환자림상자료，채용영역조조집산법제취여간염간경화림상분류유관적생물검측지표。연후，운용수증강형패협사분류기구건분류모형진행간염간경화적림상분류。결과：채용영역조조집패협사망락분류모형진행간염간경화대상성분류적정학솔위90．91％，활동성분류정학솔위94．09％，이사용BP신경망락적대상성분류정학솔위76．82％，활동성분류위85．45％。결론：영역조조집패협사망락분류방법가이유효지진행간염간경화림상분류，병능구위림상의학진단연구제공삼고。
Objective:A kind of Data mining model by Bayesian Network Based on Neighborhood Rough Sets was put forward to research the means of clinical classification of post-hepatitic cirrhosis .Methods:Ac-cording to the clinical data of 355 patients with post-hepatitic cirrhosis ,Neighborhood Rough Sets was made use of feature selection ,then tree augment naive Bayesian Network was applied to classifying the post-hepa-titic cirrhosis .Results :With using of Bayesian Network Based on Neighborhood Rough Sets ,the accuracy of distinguishing compensated post-hepatitic cirrhosis was 90 .91% and distinguishing active post-hepatitic cir-rhosis is 94 .09% ,it was 76 .82% and 85 .45% by that of BP Neural Network .Conclusion:Bayesian Network Based on Neighborhood Rough Sets was qualified for the classification of post-hepatitic cirrhosis ,and it may provide a useful method for the clinical research and the assessment of therapeutic effectiveness .