Based On Rough Set Data Reduction Techniques And Applied Research,

As the development of the information science, the style of the information is more and more complicated, so we need some strong analytical tool. In the early 1980s,professor Pawlak Z proposed the Rough Set Theory, this theory is a new mathematical tool which can be used to deal with imprecise, incomplete or uncertain classified questions, now it has attracted much attention of the researchers around the word.This article emphasis study on the important problem of Rough Set Theory—the reduction of the decision table, it contains attribute reduction and attribute's value reduction.The main content of this paper will be introduced in detail in the following text:First, introduce the attribute reduction algorithm which based on the discernibility matrix and the genetic algorithm, and analyze the capability of these algorithms via the application of instances. In addition, compare the attribute's value reduction algorithms which based on core's value with some heuristic value reduction algorithm.Then research the heuristic algorithms deeply, they are MIBARK, CEBARKCC, CEBARKNC and attribute frequency algorithm. As we all know, MIBARK, CEBARKCC and CEBARKNC algorithm are all based on the information measure,but they have some sharp constrats in the algorithm's starting, ending condition , complexity of time and so on. In this part, attribute frequency algorithm is improved, too. According to the result of the experimentation , the improved algorithm could gained the better reduction .In the last part of the paper, analyze the applying of the Rough set in the IDS, and do a experimentation with the kddcup99 data.