基于混合遗传算法的心脏病决策支持系统研究

Study on Medical Diagnosis Decision Support System for Heart Diseases Based on Hybrid Genetic Algorithm

将遗传算法和 BP算法相结合 ,建立了一个基于混合遗传算法的心脏病决策支持系统来鉴别诊断五种常见心脏病 (冠心病 ,高血压性心脏病 ,风湿性心脏病 ,慢性肺原性心脏病和先天性心脏病 )。一个含有 35 2份心脏病的数据库用来构建和测试了该系统。实验结果表明 ,构建的系统对这五种心脏病均有较好的诊断识别率 ,系统的平均识别准确性达 90 .6 % ,各疾病的用户准确性和程序准确性均大于 85 .0 % 。

In this study, a medical diagnosis decision support system based on hybrid genetic algorithm has been established to support the diagnosis of five common heart diseases (coronary heart disease, rheumatic valvular heart disease, hypertensive heart disease, chronic cor pulmonale and congenital heart disease). A heart disease database consisting of 352 samples was used for constructing and testing the performance of system. Cross validation of the experimental results indicate that the system we established shows high capability of classifying these five kinds of heart diseases, the mean accuracy of classification is as high as 90.6%, and the user accuracy and procedure accuracy of each disease are both above 85.0%, showing great application prospect of supporting heart diseases diagnosis in clinics.