摘要
糖尿病是一种由多种因素引起的无法根治的代谢性慢性病,而生活方式是影响糖尿病血糖变化的一个重要因素。因此,针对糖尿病患者的生活习惯,设计并实现了糖尿病演变与预测的可视分析系统,将空腹血糖划分为正常、轻度、中度与重度四个等级,构建了血糖演变图以展示单位时间内的血糖演变情况,并分析了单位时间内患者的生活习惯对于血糖的影响。为进一步分析影响程度,采用了不同的机器学习算法预测当前生活习惯下的血糖变化。设计可视化图形提供多视图,通过多角度的展示与交互分析来挖掘不同血糖变化的生活习惯差异,并对各类机器学习模型的预测效果进行分析与评估。便于糖尿病患者发现生活习惯的不足,改善生活习惯,减小血糖严重的可能。最后,通过案例分析进行实验,验证了该系统的有效性。
Diabetes is an incurable metabolic chronic disease caused by many factors,and lifestyle is an important factor that affects the changes in diabetes blood sugar.Therefore,a visual analysis system for diabetes evolution and prediction is designed and implemented for the living habits of diabetic patients.The fasting blood glucose is divided into four levels:normal,mild,moderate,and severe.A blood sugar evolution map is constructed to show the evolution of blood sugar per unit time,and the impact of the patient’s living habits on blood sugar per unit time is analyzed.In order to further analyze the degree of impact,different machine learning algorithms are used to predict blood glucose changes under current lifestyle habits.The visualization graphics is designed to provide multi-view,and according to multi-angle display and interactive analysis,the differences in living habits of different blood sugar changes are explored,and the prediction effects of various machine learning models are analyzed and evaluated.It is convenient for diabetic patients to find the lack of living habits,improve living habits and reduce the possibility of serious blood sugar.Finally,experiments are conducted through case analysis to verify the effectiveness of the system.
作者
母东生
蔡梦杰
王桂娟
陈华容
吴亚东
MU Dong-sheng;CAI Meng-jie;WANG Gui-juan;CHEN Hua-rong;WU Ya-dong(School of Computer Science and Technology,Southwest University of Science and Technology,Mianyang 621000,China;School of Computer Science and Engineering,Sichuan University of Science&Engineering,Zigong 643000,China)
出处
《计算机技术与发展》
2021年第10期154-160,共7页
Computer Technology and Development
基金
国家自然科学基金(61872304)
国防科研项目(19zg6106,19zg6105)
四川省杰出青年基金(19JCQN0108)
四川轻化工大学人才引进项目(2020RC20)
四川省科技计划项目资助(2020YFG0151)。
关键词
糖尿病
血糖演变
预测模型
可视化
可视分析
diabetes
blood glucose evolution
predictive model
visualization
visual analysis
