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

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 high-quality experimental data, in order to process and analyze.