基于K-means聚类分析算法的2型糖尿病动态血糖监测数据分析  

Analysis for monitoring data of type 2 diabetes mellitus based on K-means algorithm

作　　者：韦哲[1,2] 吕克难 王能才[1] 

机构地区：[1]兰州军区兰州总医院信息科,甘肃兰州730050 [2]兰州理工大学电信学院,甘肃兰州730050

出　　处：《中国医学装备》2016年第11期13-16,共4页China Medical Equipment

摘　　要：目的:探讨分析基于K-means聚类分析算法的2型糖尿病动态血糖监测数据,以解决动态血糖测量仪所测数据中的噪声和干扰信号问题,得到适用于灰色关联度分析法的实验数据。方法:引入K-means聚类分析算法处理和分析由动态血糖仪测得的糖尿病患者60 min血糖值的数据,去除误差较大的数据点,使平均数值更加可靠。结果:K-means聚类分析算法对生成所需的、无干扰地对患者60 min内间隔5 min的血糖值实验数据进行处理,并与采用K-means分析算法处理之前的数据进行对比。结论:K-means聚类分析法能够有效去除干扰和噪声信号,获得高质量的实验数据,有利于对动态血糖监测数据进行处理和分析。Objective： To analyze the monitoring data of type 2 diabetes mellitus based on K-means algorithm to avoid noise and interference signals in glycemic measurement and get experimental data applicable to Gray Relational Method. Methods： We use the data of a patient who named Mr. Li from the information department of one tertiary referral hospital in Lanzhou which includes course note of disease and his health record. And we use K-means algorithm to process and analyze his glycemic data in 60 minutes to remove error data point. Results： We can get Mr. Li＇s necessary and undisturbed experimental data in 60 minutes. Conclusion： K-means algorithm holds a higher efficiency in removing noise and interference signals to obtain highquality experimental data, in order to process and analyze.

关 键 词：血糖监测 数据处理 K-means聚类分析 灰色关联度分析法 

分 类 号：R587.1[医药卫生—内分泌]