Research And Application Of Rough Sets-Attribute Reduction On Power Market

Rough Sets (RS) theory was put forward innovatively by PawlakZ. of Polish scholars in 1982,which can effectively express uncertain or imprecise knowledge, acquire knowledge from the data well and reason by uncertainty, incomplete knowledge, not depending on additional information. As we know, usually there are redundant data, missing data, uncertain data and inconsistent data in the databases, which have become a great barrier to extracting knowledge from databases. Therefore, the model of knowledge discovery based on Rough Set plays an important role in information systems.Firstly, the domestic and international present research situation and the theory and application of the results are introduced. Secondly, the basic concepts of rough set theory and the concepts which are used continuous attribute discretization and attribute reduction are introduced. And then in this paper continuous attribute discretization, used to data pretreatment, and attribute reduction algorithm, used to data mining, are detailed. On analyzing research results of predecessor's basis, an improved Boolean logic attribute reduction algorithm is given. From the perspective of reducing redundant, the improved algorithm distinguishes properties with the attribute importance, and selected candidate attribute breakpoint according to attributes importance. In addition, from the view of attributes importance and incompatible decision table, combining the attribute importance ideas and solutions, we improve the attribute reduction algorithm based on Skowron discernibility matrix. The experiment results illustrate that these improved algorithms are effective. Last, the improved algorithm is used in the power market data mining decision support system, to help decision-makers find customers from the customer base library and analyze their difference, and then to achieve decision-making purposes.