| Nowadays, the society has already entered the network information ages. The computer and networks fly to develop to make rapid increment of the whole data of realm and information, and because of the mankind participation the indetermination of the data and the information systems becomes more obvious. How to dig out latent and useful data from large, chaotic data puts forward the unprecedented challenge to the mankind. Thus a brand-new realm of the artificial intelligence research is produced-Data Mining(DM). Currently, commonly used technology of Data Mining is: Statistical Analysis Method, Artificial Neural Networks, Decision Tree, Genetic Algorithms etc.. Among so many methods, Rough Set theory is a kind of more valid method to deal with the complicated system. Rough Set theory which has developed in recent years is a new method for analyzing and dealing with uncertain and incomplete data. The theory is a kind of data analysis theory which was put forward by Pole Z.Pawlak in 1982. Its main thought is: keeping the constant premise of the information system classification ability, the problem's classification or decision rule is derived by knowledge reduction. The application of Rough Set in Data Mining has obviously superiority-it doesn't need to provide any knowledge outside of the data which needs to be processed, makes use of the equal value relations to measure the indetermination degree of knowledge, and so avoids to the error which is brought about by subjective evaluation of knowledge. Exactly because of this, Rough Set is successfully applied in Machine Learning,Artificial Intelligence, Pattern Recognition ,Intelligence Information Processing etc.. Currently, although people in domestic have certain understanding of the theory, the theory's application in Data Mining still has some shortage. An effective and viable algorithm hasn't put forward yet. If ever, it also can't carry on the processing nicely, and has certain weakness. This obstructs the theory's application greatly.
|
PDF Full Text Request