Based On Rough Set Attribute Reduction Algorithm Research

Rough set theory is a new mathematical tool to process the fuzzy and uncertainty knowledge.The main idea is to develop the decision-making rules or classification,under the premise of keeping the ability of classification and adoption of knowledge reduction. Attribute reduction is one of key problems in the theoretical research of Rough Set, to design the effective algorithm is very important for the application of Rough Set theory. The article designs two different attribute algorithms based on Boolean Matrix and the significance of attribute.At present, all of the concepts and computing about Rough Set theory in the decision-making table are generally in the forms of algebra expressed, attribute reduction about Rough Set is studied in this form. In the form of algebra, a lot of theories and concepts about Rough Set of have less intuition, and some of the heuristic algorithm has large amount, and most of it is the non-complete algorithm. Professor Li demonstrated that attribute reduction is equivalent in the two different forms of the Boolean matrix and algebra,but Boolean matrix is more intuitive and can reduce the storage space. On this basis, new method of complete data reduction based on Boolean matrix is put forward. In order to get the minimum reduction in information system, this paper adds the converse delete action until cannot delete, this guarantee the completeness of the algorithm. We give a demonstration at last section of the paper, and it verify the improved method validity.In the decision-making table, the property is not as important as the different attributes of different significance, so there is a need to study the significance of the attributes. A attribute reduction algorithm is designed based on restrictive positive region and attributes significance. This article first translates attribute reduction in decision-making table into attribute reduction in the decision-making to simplify the list, followed redefines the concept of restrictive positive region, and then get some nature about it. Based on restrictive positive region, a new and relatively reasonable algorithm is designed for reducing searching space as quickly as possible. In other words, it makes uses of the restrictive positive region to reduce the range of dealing dates by appending the most significance of attributes to core of attributes from original set of core attributes, which can descend the complexity of algorithm. Finally, an example is used to illustrate the efficiency of the algorithm. What is more, the algorithm is also suitable for inconsistent decision table.