Research And Application On Attribute Reduction Methods Based On Information Entropy

Multiple-attribute decision-making (MADM) methodology has been one of the top issues the researchers discussed, which has been successfully applied in many real-life problems in engineering, finances, market analysis, management and others. We analyzes the characteristic of the complex decision methods and introduce international advanced thoughts and methods such as rought sets theory into classical mufti-attribute decision .Rough set is a new method of data mining. Its basic theory is through attribution reduction, obtaining knowledge with the same ability of classification. Attribute reduction based on information entropy is one of the important issues of rough set theory.In this thesis, by the comprehension and analysis of data mining algorithm based on the rough set theory, several heuristic algorithms for attribution reduction based on knowledge entropy have been proposed. After analyzing the algorithm, the reason for incomplete reduction is found. Regarding the significance of attribute defined from the viewpoint of information theory as heuristic information the construction of the heuristic information is discussed, and an improved algorithm which combined the algebra of the rough set theory and the conditional information entropy is put forward. The conditional information entropy is the heuristic information of the attribute reduction. Algorithm is based on conditional information entropy for reduction of decision table, but the algorithm is incomplete for some decision table, so an improved algorithm which combined the algebra of the rough set theory and the conditional information entropy is put forward. The experiment results show that this algorithm can find the minimal reduction for decision table.The disadvantages of the current conditional information entropy are analyzed.So extensive attention has been given .Based on this entropy the new significance of an attribute is defined and compared with two significances of this attribute based on the attribute is defined and compared with two significances of this attribute based on the positive region and the current conditional information entropy respectively. Finally, a heuristic algorithm for knowledge reduction is designed and an efficient algorithm for computing conditional information entropy is proposed. Also, this reduction algorithm is more capable of finding the minimal or optimal reductSince rough set has been presented its theory and method get development continuously. Furthermore we put forward a set of mufti-attribute decision methods based on rough set theory. We prove these sets of method have the theory value and practice value.