Studies On Tremendous Amounts Of Data Oriented Rough Set Theory And Method

Since rough set has been presented its theory and method get development continuely. In some aspects rough set overcomes many deficiencies in traditional data analysis, so it has been given extensive attention at home and abroad. The dissertation makes academic research on rough set theory for tremendous amounts of data, and gives a comprehensive comment about the research development. The emphases comprise reduct analysis, variable precision rough set model, dynamic reduct, decision analysis of rule set, redundant knowledge decision fusion and so on.For the reduct analysis based on the merging, it shows the most essential property and the subdivision hierarchy relation about attributes reduct, that is partially ordered lattice construction. Then the impact of condition class merging to consistence of decision table is analyzed.According to the properties of partition the information entropy, the decision entropy and the condition entropy are studied. The properties and interrelations among them are described as well. Then the impacts on uncertainty and the reduct in decision information system are analyzed. For time decision information system the emphases are acquirement and reduct, and the time significance reduct strategy is put up.By introducing the inclusion proportion threshold value for each condition class, it makes a lucubration on reduct anomaly about variable precision rough set model, and describes the range relation between inclusion proportion and quality of classification. Then it analyzes the reduct anomaly when inclusion degree vibrates and positive area changes. The basic ideas are presented to eliminate reduct anomaly. At the end it gives the range reduct definition, and realizes the range reduct algorithm. All of this develops the reduct of the variable precision rough set model.The dissertation describes the model of the dynamic reduct and discusses various formal dynamic reducts. It presents the new method for F family computation, and the precision coefficient is introduced to sampling problem. A complete dynamic reduct framework isconstructed, and the property analysis is made in detail.This thesis firstly presents the dynamic core concept according to the dynamic reduct model. Then it describes multi-hierarchies formal definition of dynamic core, and discusses some properties of dynamic core. Especially it proves that the intersection of dynamic reduct comprises dynamic core, which means that dynamic core has the essential character about feature core.The article describes many metrics of decision rules and analyzes the properties for these metrics. Then it presents the metrics for rule set, which shows the properties of a rule set in general. All of them play an important role for decision of redundant knowledge. For the deficiency of attribute values it presents the rule discernibility matrix, which shows the properties of a rule set in general. All of them realize attribute values reduct and play an important role for decision analysis.Based on basic theory of model integration, a formalization representation of model is given. The thesis also presents compound model relation and model integration method. Each knowledge-base can be viewed as a single decision model, then decision fusion is realized by model integration. It achieves a comprehensive decision at model level. Additionally, the existence of model integration is analyzed in detail and several sufficient conditions are proved.