Research On Key Technologies Of Decomposition Of Decision Information System

Rough set theory is an approach of data analysis dealing with imprecise, inconsistent and incomplete information. The data analysis and processing technology based on rough set theory have been widely applied. Decomposition of decision table is an effective tool to deal with large decision table with great quantity and high data complexity, and it can be valid to improve the efficiency of data analysis. Based on rough set theory, the thesis mainly focuses on the research of key technologies in decomposition.The concepts and properties of rough set theory and decomposition are described, and the methods, principles and features are also discussed in details. Some deficiencies in the attribute set decomposition based on attribute significance are discussed. The differences between the quality of approximation classification and the attribute significance to decision making are analyzed. With the classification and the attribute significance, the relationships between condition attributes and decision attributes are considered to increase the attribute classification rate in the decision table, and a new attribute selection criterion is proposed. At last the decision table is decomposed with the attribute selection criterion.Decision rule generalization may appear during attribute set decomposition, so it may lead to differences of the rules from the original decision table and the child decision tables. The equivalence in attribute set decomposition is deeply studied. From the view of keeping the equivalence in decomposition of decision table and improving the classification rate in the child decision tables, novel attribute set decomposition approaches are proposed.Through simulation and comparing with existing methods of attribute set decomposition, the results of experimental illustrate that the algorithms are more effective and accurate.