人工神经网络应用于糖尿病/糖耐量异常的疾病分类研究

钱玲; 施侣元; 程茂金

Abstract

钱玲,施侣元,程茂金.人工神经网络应用于糖尿病/糖耐量异常的疾病分类研究[J].Chinese journal of Epidemiology,2003,24(11):1052-1056

人工神经网络应用于糖尿病/糖耐量异常的疾病分类研究

Study on the application of artificial neural network on diabetes mellitus/insulin-glucose tolerance classification

Received:January 16, 2003

DOI：

KeyWord: 人工神经网络学习向量量化网络糖尿病糖耐量异常疾病分类

English Key Word: Artificial neural network Learning vector quantization neural network Diabetes mellitus insulin glucose tolerance Classification of disease

FundProject:

Author Name	Affiliation
QIAN Ling	National Healtheducation Institu te, Chinese Center for Disease Control and Prevention, Beijing 100011, China
SHI Lv-yuan	华中科技大学同济医学院公共卫生学院
CHENG Mao-jin	华中科技大学同济医学院公共卫生学院

Hits: 3512

Download times: 1188

Abstract:

目的探讨人工神经网络(ANN)用于疾病分类研究的前景。方法利用某矿区1996年糖尿病现况调查资料,采用学习向量量化(LVQ)网络和判别分析方法进行糖尿病/糖耐量(DMIGT)异常正常状态的判别比较;同时人为设置变量缺损值,检验LVQ网络对缺失数据的适应性。结果LVQ网络结构为25→13→3;网络判断准确率为96.98%,对血糖异常者的正确判断率为92.45%。利用逐步判别分析建立的含11个变量的判别方程的判断准确率为87.34%,对血糖异常者的正确判断率为85.53%。LVQ网络对带缺失项样本的误判比例为130,判别分析则为730。结论利用LVQ网络进行疾病分类预测,不仅能获得更好的预测效果,而且对资料的类型、分布不作任何限制,也不需要对分析变量做任何处理,还能很好地处理带缺失项的资料,是一种很好的流行病学分类预测新方法

English Abstract:

Objective To discuss the potentialapplication of artificial neuralnetwork( ANN)on the epidemiological classification of disease. Methods Learning vector quantization neural network ( LVQNN) and discriminate analysis were applied to data from epidemiological survey in a mine in 1996. Results The structure of LVQNN was 25※13 ※3.The total veracity rates was 96. 98%, and 92. 45% among the abnormal blood glucose individuals. Through stepwise discriminate analysis, the discriminate equations were established including 11 variables with a total veracity rate of 87. 34%, buTwas 85. 53% in the abnormal blood glucose individuals. Further analysis on 30 cases with missing values showed thaTthe disagreemenTratio of LVQ was 1 30,lower than thaTof discriminate analysis of 730. Conclusions Compared to the conventional statistics method, LVQ noTonly showed better prediction precision, buTcould treaTdata with missing values satisfactorily plus iThad no limiTto the type or distribution of relevanTdata, thus provided a newpowerful method to epidemiologic prediction.

View Fulltext Html FullText View/Add Comment Download reader