三角模糊不完备三支群决策及其在糖尿病诊断中的应用  

Triangular Fuzzy Incomplete Three-Way Group Decision-Making and Its Application in Diabetes Diagnosis

作　　者：王安娜 张超[1,2] 王丽茵 程钰婷 范雪晴 林铭炜 WANG Anna;ZHANG Chao;WANG Liyin;CHENG Yuting;FAN Xueqing;LIN Mingwei(School of Computer and Information Technology,Shanxi University,Taiyuan 030006,China;Key Laboratory of Computational Intelligence and Chinese Information Processing of Ministry of Education,Shanxi University,Taiyuan 030006,China;College of Computer and Cyber Security,Fujian Normal University,Fuzhou 350117,China)

机构地区：[1]山西大学计算机与信息技术学院,山西太原030006 [2]山西大学计算智能与中文信息处理教育部重点实验室,山西太原030006 [3]福建师范大学计算机与网络空间安全学院,福建福州350117

出　　处：《福建师范大学学报（自然科学版）》2024年第5期1-16,共16页Journal of Fujian Normal University：Natural Science Edition

基　　金：国家自然科学基金资助项目(62272284、62072294、61972238);中央引导地方科技发展资金项目(YDZJSX2024D015);福建省“雏鹰计划”青年拔尖人才计划(F21E0011202B01);山西省科技创新青年人才团队项目(202204051001015);山西省高等学校优秀成果培育项目(2019SK036);国家大学生创新创业培训计划(20230014);山西大学第21届大学生创新创业培训计划(202210108012);山西大学文瀛青年学者。

摘　　要：为同时应对不确定信息表示与风险信息融合对群决策带来的挑战,构建一种三角模糊不完备三支群决策方法,并将其应用于糖尿病诊断决策。首先,针对信息不确定性蕴含的模糊性和不完备性,分别引入三角模糊集和不完备信息系统的概念。通过与多粒度三支决策结合,构建了可调多粒度三角模糊概率粗糙集模型。然后,根据离差最大化法计算属性权重与专家权重,结合ELECTRE(elimination et choice translating reality)方法建立了三角模糊多属性群决策方法。最后,通过对糖尿病患者数据的案例分析和评估,验证了所提方法的可行性和有效性。该方法不仅从不确定信息表示、风险信息融合和最优粒度选择的视角丰富了多粒度三支群决策理论,而且推动了糖尿病智能诊断方面的应用。This paper constructs a triangular fuzzy incomplete three-way group decision-making method to simultaneously address the challenges posed by uncertain information representation and risk information fusion in group decision-making,and applies this method to diabetes diagnosis.First,the concepts of triangular fuzzy sets and incomplete information systems are introduced to address the fuzziness and incompleteness associated with information uncertainty.By integrating these concepts with multi-granularity three-way decision-making,an adjustable multi-granularity triangular fuzzy probabilistic rough set model is developed.Subsequently,using the maximum deviation method,attribute weights and expert weights are calculated,and combined with the ELECTRE(elimination et choice translating reality)method to establish a triangular fuzzy multi-attribute group decision-making method.Finally,the feasibility and effectiveness of the proposed method are validated through case analysis and evaluation of diabetic patient data.This method not only enriches the theory of multi-granularity three-way group decision-making from the perspectives of uncertain information representation,risk information fusion,and optimal granularity selection,but also advances its application in intelligent diabetes diagnosis.

关 键 词：多粒度三支群决策 三角模糊集 概率粗糙集 不完备信息系统 糖尿病诊断 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]