摘要
近年来,糖尿病发病率愈发升高,糖尿病的初期诊断也变得愈发复杂。糖尿病病因和多种因素相关,具有一定的不确定性,基于此,提出一种利用机器学习(例如支持向量机SVM)来对糖尿病诊断进行辅助建模的方法,选取美国国家糖尿病数据组提供的糖尿病患者诊断数据(共计231项),对SVM和利用灰狼以及布谷鸟算法对其优化后的结果进行比较,实验结果表明:传统算法模型精度较低,而优化后精度都有所提升,且布谷鸟优化算法收敛较快,寻优能力最强,平均精度达到78%以上,用于糖尿病实际数据诊断建模最佳。
In recent years,the incidence of diabetes has been increasing,and the initial diagnosis of diabetes has become more complex.Diabetes cause and a variety of factors related,has certain uncertainty,based on this,this paper presents a use of machine learning,such as support vector machines SVM to diabetes diagnosis method for auxiliary modeling,the author selected the national diabetes diabetes diagnosis data from a data set(231),the algorithm of SVM and using the gray Wolf and cuckoo comparing the optimized result,the experimental results show that the traditional model precision is low,and the accuracy are improved after optimization,and cuckoo optimization algorithm converges faster,optimization,the most powerful average accuracy of 78%or more,It is best used for diagnostic modeling of actual data of diabetes.
作者
裴修侗
周晓东
陈凯祥
乐洋
PEI Xiudong;ZHOU Xiaodong;CHEN Kaixiang;LE Yang(Nanjing University of Posts and Telecommunications, Nanjing 210023, China;Smart Health Big Data Analysis and Location Services Engineering Laboratory of Jiangsu Province, Nanjing 210023, China)
出处
《技术与市场》
2021年第8期9-11,共3页
Technology and Market
基金
南京邮电大学教学改革研究重点招标课题(JG03219JX64)
南京邮电大学校级科研基金资助项目(NY219068)
江苏省智慧健康实验室开放课题项目(SHEL219002)。
关键词
支持向量机
糖尿病诊断
布谷鸟算法
灰狼算法
support vector machine(SVM)
diagnose of diabetes
Cuckoo algorithm(CS)
Gray Wolf algorithm(GWO)
