Research On Approximate Description And Measure Of Uncertain Concept In Multi-Granularity Spaces

With the progress of human society and the rapid development of science technology,the data in production and life are surging all the time.How to deal with uncertain data and mine valuable knowledge intelligently and efficiently is a key problem in knowledge discovery at present.Rough set theory,introducing two precise sets to approximate the uncertain concepts,provides a novel perspective of processing uncertain data in multigranularity spaces.Based on this,considering both positive and negative information simultaneously,vague set theory has been well applied in research on rough approximate spaces.Therefore,rough set theory and vague set theory have been the key roles for approximate description and measure of uncertain concepts in multi-granularity spaces.In order to better describe the uncertain problem,this thesis studies approximate description of the uncertain concept with different granularity layers in multi-granularity spaces and discusses the rules of uncertainty with changing granularities.The main researches are presented as follows:In order to study the approximate description of uncertain concepts,the existing approximation set models of rough sets are analyzed firstly.And the optimal approximation set of rough sets is proposed to describe the uncertain concept from the point of similarity.Then,when the uncertain concepts have vague characteristics,the uncertainty analysis of the existing approximation set models of vague sets is carried out.Next,improved stepvague set is proposed through local optimum approaching global optimum.Finally,it is applied to three-way decisions which provides a good theoretical basis for intelligent decision-making based on the uncertain concepts.In view of the measure and analysis of uncertain concepts with changing granularities in multi-granularity spaces,firstly,the relationships between the uncertainty of vague value and its internal parameters are analyzed.Then change rules of uncertainty of approximate concept in vague sets with changing granularities are discussed.Next,from the angle of three decision-making,change rules of the uncertainty of the lower approximate concepts with changing granularities in the pessimistic decision-making are discussed.Finally,and the change rules of the uncertainty of boundary regions with changing granularities in the sequential three-way decisions are discussed.It provides the theoretical basis for the dynamic intelligent decision-making in the multi-granularity spaces in various scenarios.To sum up,this thesis studies the approximate description of uncertainty concepts through rough sets and vague sets from the perspective of uncertainty,and then analyzes the variation trend for uncertainty of approximate concepts with changing granularities,which provide theoretical basis on granularity selection for intelligent decision-making in multi-granularity spaces.