Research On Data Reduction Based On Rough Set

After we studied and summarized the general development, advances and challenges of rough set and knowledge discovery, this paper proposes theory of rough set, data reduction and the extended rough set model. In addition, we study the knowledge discovery based on rough set, and we propose a data mining model based on clustering and rough set. Finally, we study the synthetic information management based on rough set.The detail results follows:1 .We study the basic knowledge of rough set, the method of data reduction based on rough set and the extended rough set models, and then introduce the basic course of knowledge discovery, the method of stuffing the missing attribute value and the technology of discretization.2. We propose a data mining model based on clustering and rough set. This model reduces the original data with clustering algorithm firstly, and eliminates the suspicious information, which makes the data consistent. Then the data is made a qualitative analysis and reduced based on rough set theory. The data is reduced in both horizontal and vertical directions by using hierarchical clustering and rough set methods.3. We study the synthetic information management based on rough set. We propose a method of data mining that can deal with the information with the time sequences data. Firstly, the pattern of time sequences data was discovered by using the similarity search on time sequences. Secondly, the data of time sequences was transformed into the disperse data. Lastly, the data was reduced by using the rough set theory. The rules of information containing the time sequences data were discovered by using this data mining model. We propose an algorithm that integrates the rough set theory with alternative covering neural networks.