The Research On The Reduction And Its Application Of Rough Sets

Facing the notorious data, we need a kind of efficient tool(s) to find some useful knowledge. Rough sets is one of such tools. Finding the reductions and the core(s) of Information System plays an important role in Rough Sets. We may get reductions through some methods. But the key of the question is how to delete the redundancy attributes by the discern function. On the other hand, we need some good rules, for example, the Minimum Description Length, which supervise our action to be more efficient.Therefore, we put focus on how to get reductions.First some methods are introduced about the relative reductions in Decision Information System and in uncompleted system. For example, by the method of the Binary string and the 'and' of logic operation, we may get reductions which need not create the discern matrix and avoid using the absorption law of Boolean algebra; by defining some new functions, we may also get reductions. The parts of the methods are programmed by MATLAB.Then the rules may be got in the new system after reducing. We put them into IF-THEN and Decision Trees by some methods. We also discuss how to learn in rule sets. Last, we gave some applications on Rough Sets that demonstrate the tool of Rough Sets is very important in KDD.