摘要
糖尿病被称为现代疾病中的“第二杀手”,开发糖尿病的早期检测方法至关重要。本研究基于一个含320个早期糖尿病样本和200个健康样本的早期糖尿病数据集,通过主成分分析(Principal Component Analysis,PCA)提取重要特征,构建并训练用于预测早期糖尿病的模糊宽度学习系统(Fuzzy Broad Learning System,FBLS)模型。FBLS模型在该早期糖尿病数据集上的准确率、灵敏度、特异性、F1-Score和受试者工作特征曲线下面积(Area Under the Receiver Operating Characteristic Curve,AUC)分别为85.38%、79.69%、94.5%、95.86%、0.7967。该结果表明FBLS模型具有较好的早期糖尿病风险预测性能,可以实现糖尿病的早期风险预测。
Diabetes is known as the"second killer"of modern diseases.It is very important to develop early detection methods for diabetes.This research is based on a data set of early diabetes with 320 samples of early diabetes and 200 health samples.Through Principal Component Analysis(PCA),important features are extracted,and a Fuzzy Broad Learning System(FBLS)model for predicting early diabetes is constructed and trained.The accuracy,sensitivity,specificity,F1-Score and Area Under the Receiver Operating Characteristic Curve(AUC)of the FBLS model on the early diabetes dataset were 85.38%,79.69%,94.5%,95.86%and 0.7967,respectively.The results show that the FBLS model has good risk prediction performance for early diabetes,and can realize early risk prediction for diabetes.
作者
谢双波
Xie Shuangbo(Intelligent Manufacturing College of Hunan University of Science and Engineering,Yongzhou,China)
出处
《科学技术创新》
2025年第18期64-67,共4页
Scientific and Technological Innovation
基金
湖南省教育厅科研项目-一般项目(24C0477)
永州市指导性科技计划项目-一般项目(2024YZ018)
湖南科技学院校级科研项目(23XKYZZ10)。
关键词
早期糖尿病
模糊宽度学习系统
主成分分析
early diabetes
fuzzy broad learning system
principal component analysis
