基于BP神经网络的脑卒中患者分类方法研究  

Research on classification method of patients with cerebral apoplexy based on BP neural network

作　　者：吴洁[1] 曲畅[2] 王华[3] Wu Jie Qu Chang Wang Hua(Electrical and Information Engineering College of Beihua University, Jilin 132012, China Electronics and Information Engineering College of Changchun University of Science and Technology, Changchun 130022, China Computer Application Department of Shanghai Technical Institute of Electronics ＆ Information, Shanghai 201411, China)

机构地区：[1]北华大学电气信息工程学院,吉林吉林132012 [2]长春理工大学电子信息工程学院,吉林长春130022 [3]上海电子信息职业技术学院计算机应用系,上海201411

出　　处：《无线互联科技》2016年第24期140-142,共3页Wireless Internet Technology

基　　金：吉林市科技计划项目;项目编号:201537112

摘　　要：人工神经网络的应用逐渐广泛,已经发展到各个前沿领域,在神经专家系统、模式识别、智能控制、组合优化、预测等领域得到成功应用。文章通过对经过PCA算法降维预处理后的数据进行BP神经网络数据分类,比较和分析数据经BP神经网络、RBF神经网络、ANFIS神经网络后的分类误差。经实验分析,BP神经网络、RBF神经网络、ANFIS神经网络分类误差分别为0.000 5,971.935,0.025 3,传统BP神经网络更适合文章中所采集的脑卒中患者数据分类。The application of artificial neural network is gradually widely, which has developed to the frontier fields such as the neural expert system, pattern recognition, intelligent control, combination optimization, forecasting and other fields. This article mainly used the data after PCA algorithm dimension reduction pretreatment. Comparing and analyzing the data classification error though the BP neural network, RBF neural network and ANFIS fuzzy neural network. Through experimental analysis, the classification error respectively is 0.000 5, 971.935, 0.0 253 by BP neural network, RBF neural network and ANFIS neural network, traditional BP neural network is more suitable for data classification of cerebral apoplexy patients collected in this paper.

关 键 词：BP神经网络 脑卒中 分类方法 信息处理 

分 类 号：R743.3[医药卫生—神经病学与精神病学] TP183[医药卫生—临床医学]