The Study Of Attribute Reduction Method Based On Core Samples Set

As development of the database technology and application of database management system, huge amounts of information is stored up in different forms of data. Data mining is a kind of data processing technology to get valuable and developed information from these data. The attributes of these data are not equally important, even some attributes are redundant. Therefore, refining attribute sets of database (attribute reduction) become an important link of data mining technology. Based on the formal representation of rule knowledge, this paper put forward the concept of determinacy knowledge and core samples set, then establish the attribute reduction method and the computing core attributes algorithm based on core samples set combining with decision tree algorithm.Firstly, this paper first analyzes the essence characteristics of decision tree algorithm. Based on the formal representation of rule knowledge, we put forward the concept of determinacy knowledge and core samples set and prove the invariance of the rule knowledge;Secondly, combining with decision tree algorithm, we establish the attribute reduction method and the computing core attributes algorithm based on core samples set;Finally, we reveal that the purity of the envoys point is not 1 because of compulsory fitting of decision tree in incoordinate environment. Through this analysis, we we put forward the concept ofβ-determinacy knowledge andβ-core samples set contrastively and establish the attribute reduction method based onβ-core samples set.Combining with instantiation, we analyzes the features of the performance of these methods. The results show that they all have strong maneuverability and can deal with different types of database reduction effectively.