摘要
多粒度三支决策隐私保护融合粒计算与三支决策理论,实现对数据的多视角、多层次分析与处理,在保障隐私的同时提升数据可用性。综述了该领域的研究进展,重点从多粒度隐私保护模型、多粒度三支决策隐私保护模型及多粒度序贯三支决策隐私保护模型3个方面,系统探讨了粒计算与三支决策在数据匿名化和差分隐私技术中的应用与发展,并分析了不同多粒度隐私保护模型的关系。尽管该方法在平衡数据可用性与隐私性方面表现出优势,但仍面临高维数据处理复杂性、动态数据流实时匿名化及个性化隐私需求调控等挑战。未来研究可聚焦自适应多粒度隐私保护机制、跨领域数据协同保护,以及结合机器学习与差分隐私的智能化隐私防护策略,以提升模型的实用性与扩展性。
Multi-granularity three-way decision privacy protection integrates granular computing and three-way decision theory to enable multi-perspective,multi-level analysis and processing of data,enhancing data usability while safeguarding privacy.This paper reviews the research progress in this field,focusing on multi-granularity privacy protection models,multi-granularity three-way decision privacy protection models,and multi-granularity sequential three-way decision privacy protection models.It systematically examines the applications and developments of granular computing and three-way decision theory in data anonymization and differential privacy,and analyzes the relationships among different multi-granularity privacy protection models.Although these methods offer advantages in balancing data usability and privacy,they still face challenges such as high-dimensional data processing complexity,real-time anonymization of dynamic data streams,and personalized privacy requirement management.Future research may focus on adaptive multi-granularity privacy protection mechanisms,cross-domain data collaborative protection,and intelligent privacy-preserving strategies that combine machine learning with differential privacy to enhance the practicality and scalability of the models.
作者
钱进
杨广金
QIAN Jin;YANG Guangjin(School of Information and Software Engineering,East China Jiaotong University,Nanchang,Jiangxi 330013,China;School of Computer Science,Jiangsu University of Science and Technology,Zhenjiang,Jiangsu 212003,China)
出处
《闽南师范大学学报(自然科学版)》
2025年第4期1-18,共18页
Journal of Minnan Normal University:Natural Science
基金
国家自然科学基金项目(62466017,62066014)
江西省自然科学基金项目(20232ACB202013)。
关键词
多粒度计算
隐私保护
三支决策
数据匿名化
差分隐私
multi-granularity computing
privacy protection
three-way decisions
data anonymization
differential privacy
