基于知识粒化的粗糙集近似精度的泛化改进  

Knowledge granulation-based generalized improvement of approximate accuracy regarding rough sets

作　　者：何流 张贤勇[1,2] 唐孝 HE Liu;ZHANG Xianyong;TANG Xiao(School of Mathematical Sciences,Sichuan Normal University,Chengdu 610068,China;Institute of Intelligent Information and Quantum Information,Sichuan Normal University,Chengdu 610068,China)

机构地区：[1]四川师范大学数学科学学院,成都610068 [2]四川师范大学智能信息与量子信息研究所,成都610068

出　　处：《系统工程理论与实践》2021年第5期1343-1352,共10页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(61673285);四川省科技计划(2021YJ0085,2019YJ0529);四川省青年基金(2017JQ0046)。

摘　　要：精度是表征粗糙集概念精确性的基本度量,但具有关于知识粒化的信息缺失与量化局限.近似精度引入知识粒度进行改进.归纳分析现有五种近似精度,探讨基于知识粒化的近似精度的泛化原则与改进建模.基于实例分析,确立近似精度的"单调、一致、无关、下界"四项改进准则,验证已有近似精度的满足性;挖掘近似精度的"逼近、单调"两项量化优化,融合概念精度与知识粒度建立近似精度的泛化含参模型;提出一种改进近似精度,揭示其准则满足性与优化应用性,并用知识实例与数据实验进行有效验证.所得结果提供了近似精度的泛化构造框架与具体改进模型.Accuracy serves as a fundamental measure to quantify concept precision of rough sets,but it has information absence and quantification limitation in knowledge granulation.Approximate accuracy introduces knowledge granularity to implement improvements.Five types of existing approximate accuracy are summarized and analyzed,and generalization principles and improvement modeling are investigated based on knowledge granulation.In terms of example analyses,four types of improvement criteria(regarding monotonicity,consistency,irrelevance and lower-bound)are first established for approximate accuracy,and relevant satisfiability is inspected for existing approximate accuracy.Then,two types of quantification optimizations(regarding approximation and monotonicity)are mined for approximate accuracy,and concept accuracy and knowledge granularity are fused to set up a generalized model with parameter.Finally,a kind of approximate accuracy with improvements is proposed,its criterion satisfiability and optimization applicability are revealed,and knowledge examples and data experiments are adopted for effectiveness verification.This study provides both a generalized construction framework and a concrete improvement model for approximate accuracy.

关 键 词：粗糙集 精度 近似精度 知识粒化 四项改进准则 两项量化优化 融合含参建模 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]