Research Of Incremental Attribute Reduction Algorithm Based On Block Discernibility Matrix

The core and attribute reduction in rough set theory and its application in various fields using one of the important research content, is also one of the core issues. Attribute reduction for the given classification and decision ability of decision table under the constant remain completely, delete redundant attributes of decision table, simplifying the knowledge representation, reduced decision table processing scale, improve the efficiency of system processing, object attributes in decision table can change quickly and accurately to find the updated nuclear and attribute reduction of decision table. Most modern researchers using the attributes of decision table nuclear first, using heuristic information of algorithm of attribute reduction of decision table. Therefore, how to design a reasonable, efficient and quick decision table calculation method and attribute reduction algorithm is a very important research significance.At present, many scholars have been put forward for the decision table, and a variety of accounting method and attribute reduction algorithm and incremental algorithm of attribute reduction, however, these algorithms is mainly suitable for static decision table, when the data in the decision table, and may become inconsistent, its properties and incremental updating algorithm of attribute reduction. So to discuss and explore the object increased cases of accounting method and attribute reduction algorithm has a strong research significance.This paper first describes the research background, analyzes the research status both at home and abroad, and given knowledge of basic theory of rough set, attribute reduction of decision table is discussed systemically, and the core idea, and then in reference and concluded on the basis of predecessors’research results, made several main work as follows:(1) discuss the decision table is compressed and simplified method, and proves that the related properties of compressed simplify the decision table, using the characteristics of the classification tree, this paper proposes a decision-making table compression simplification algorithm based on classification tree.(2) Proposed by block processing way of thinking of discernibility matrix, we design a new incremental calculation method based on block difference matrix, according to different analysis of incremental objects out of four possible situation, and the corresponding processing method is given.(3) on the basis of the incremental calculation method, and puts forward the incremental attribute reduction algorithm based on block discernibility matrix, and through theoretical analysis and experiments proved the feasibility of the algorithm is efficient.