Knowledge Acquisition Of Interval-valued Information Systems Based On Similarity Of Interval Numbers

Along with the development of society and economy, people considered increasingly complexity and uncertainty of the problems, so that the data characterized things behaviors attribute is usually uncertain and some errors range, the data is referred to as interval numbers.If attribute value is interval numbers, we call it interval-valued information system. Interval numbers can effectively represent many uncertain parameters.Interval-valued information system can well consider the uncertainty of the data, and as a model for data description, it is used in wide fields of scientific research, engineering technology and decision analysis, meanwhile, effectively applied to practice. So knowledge acquisitions of interval-valued information system have received common concern from the researchers.The thesis researched the methods of knowledge acquisitions in interval-valued information system using rough set theory, and mainly included the following parts:First, for two interval numbers,the thesis defined some basic operator including intersection, union, complement and inclusion degree, and discussed their properties.By using the operators of interval numbers that have defined, the thesis constructed aαtolerance relationship and presented related concepts and properties aboutαmaximal consistent blocks under defined a tolerance relationship.Secondly, the attribute reduction is always research hotspot and core problem.The thesis constructed discernibility matrix using a maximal consistent blocks to do attribute reduction in interval-valued information system so as to simplify scale of discernibility matrix and improve computational efficiency. Furthermore,for an object in the universe,the thesis given the algorithm of relative reducts.After attribute reduction, interval-valued decision information system can acquire simplest decision rules so as to achieve the purpose of prediction. Thirdly, for multi-attribute alternative ranking, if some attribute value can not accurately be given or only be a rough description.The thesis described the kind of alternative ranking using interval-valued information system. Based on inclusion degree of interval numbers, the concept of deviation degree was proposed to structure the deviation degree matrix for alternatives, and a method for determining weight vector based on the total deviation degree optimization was also proposed.By using the proposed concept of relative closeness degree of interval number vectors the degree of an alternative close to another was depicted, and thereby alternatives were ranking.The thesis researched the methods of knowledge acquisitions in interval-valued information system based on rough set theory. Some illustrative examples were employed to validate the proposed methods and demonstrated the feasibility and practicability of the methods.