集成优化核极限学习机的冠心病无创性诊断
发布时间：2018-09-05 13:03
【摘要】：冠心病的早期无创性诊断一直是医疗诊断领域的研究热点,为了提高冠心病诊断的准确率和诊断效率,提出了一种新颖的局部Fisher判别分析(LFDA)特征提取方法和集成核极限学习机(KELM)相结合的冠心病诊断模型(LFDA-EKELM)。首先使用LFDA方法剔除不相关特征和冗余特征,找出对分类结果贡献度较高的特征子集,产生不同的训练集以训练粒子群优化的KELM分类器PSO-KELM;基于旋转森林(RF)构建集成分类器,实现冠心病的智能诊断。实验结果表明,与基于ELM、SVM和BPNN方法相比,该方法有效提高了冠心病诊断准确率、提升了诊断效率,且分类结果高于已有方法和相似方法,是一种有效冠心病诊断模型。
[Abstract]:The early noninvasive diagnosis of coronary heart disease (CHD) has been a hot topic in the field of medical diagnosis. In order to improve the accuracy and efficiency of coronary heart disease diagnosis, A novel (LFDA) feature extraction method based on local Fisher discriminant analysis and an integrated kernel limit learning machine (KELM) model for coronary heart disease diagnosis (LFDA-EKELM) are proposed. Firstly, the LFDA method is used to eliminate the irrelevant and redundant features, and the feature subsets with high contribution to the classification results are found, and different training sets are generated to train the particle swarm optimization KELM classifier PSO-KELM; based on the rotating forest (RF) to construct the integrated classifier. To realize intelligent diagnosis of coronary heart disease. The experimental results show that compared with the ELM,SVM and BPNN methods, this method can effectively improve the diagnostic accuracy of coronary heart disease and improve the diagnostic efficiency, and the classification results are higher than the existing methods and similar methods, so it is an effective diagnosis model of coronary heart disease.
【作者单位】： 深圳信息职业技术学院数字媒体学院;
【基金】：国家自然科学青年基金资助项目(61303113) 广东省自然科学基金资助项目(2016A0303100072) 深圳市科技计划资助项目(GJHZ20150316112246318)
【分类号】：R541.4;TP181
本文编号：2224360
[Abstract]:The early noninvasive diagnosis of coronary heart disease (CHD) has been a hot topic in the field of medical diagnosis. In order to improve the accuracy and efficiency of coronary heart disease diagnosis, A novel (LFDA) feature extraction method based on local Fisher discriminant analysis and an integrated kernel limit learning machine (KELM) model for coronary heart disease diagnosis (LFDA-EKELM) are proposed. Firstly, the LFDA method is used to eliminate the irrelevant and redundant features, and the feature subsets with high contribution to the classification results are found, and different training sets are generated to train the particle swarm optimization KELM classifier PSO-KELM; based on the rotating forest (RF) to construct the integrated classifier. To realize intelligent diagnosis of coronary heart disease. The experimental results show that compared with the ELM,SVM and BPNN methods, this method can effectively improve the diagnostic accuracy of coronary heart disease and improve the diagnostic efficiency, and the classification results are higher than the existing methods and similar methods, so it is an effective diagnosis model of coronary heart disease.
【作者单位】： 深圳信息职业技术学院数字媒体学院;
【基金】：国家自然科学青年基金资助项目(61303113) 广东省自然科学基金资助项目(2016A0303100072) 深圳市科技计划资助项目(GJHZ20150316112246318)
【分类号】：R541.4;TP181
【相似文献】
相关期刊论文 前1条
1 邵耕;应用无创性诊断方法的一些问题[J];中华心血管病杂志;1994年03期
，本文编号：2224360
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2224360.html
