Knowledge Reduction Based On Extended Rough Set Model In Set-valued Information Systems

Knowledge reduction for set-valued information system are studied based on the α tolerancerelation,with rough set theory as tool, set-valued information system as the object and knowl-edge reduction as objective. Set-valued information systems with a wider application, whichintroduces the default and multiple attribute value to the information system all together, isa generalization for the general information system. Therefore, it is of significance to studythe rough set model and its attribute reduction of set-valued information systems. The maincontent of this text are as follows:In view of the α tolerance relation, the text put forward the definition, the judgment andα level discernibility matrix of the knowledge reduction in set-valued information systems. Bymeans of which, the text summarizes the detailed operation methods for knowledge reductionand discusses the efect of tolerance grade upon the knowledge reduction finally.Introducing the α tolerance relation to the set-valued decision information systems, thetext advance the α level discernibility matrix under the α tolerance relation and obtain thespecific operation methods to consistent or inconsistent set-valued decision information systems.By the help of applying the evidence theory, a kind of efective uncertain reasoning methodin the information fusion technology, to the set-value information systems, the text define thebasic probability distribution function based on α tolerance relation and bring forward themethods of knowledge reduction. And the text illustrate the validity of this method throughthe concrete examples finally.