The Designing And Implement Of The Data Classifier In Data Mining

With an amount of data in various databases,it is important to find a way to discover usable information from databases efficiently and effectively. Data mining technology is an efficient solution to this problem,it has been paid widely attention to by science,industry and business. Data classification is one of the data mining technologies,previous studies on data classification have been focused on tuple-oriented method,but the methods developed must repeatedly examine data tuple by tuple,which leads to quite inefficient algorithms for use with large databases .Howeber,Concept-Based Attribute-Oriented Imdictopm will reduce the count of scanning tuple and improve the efficiency of data classification .This paper discusses the relevant problems of applying the method of Concept-Based Attribute-Oriented to find classification rule,also expounds the implemented data classifier. Firstly,we overview the recent development of data mining and data classification. Secondly,we analyse general criteria of data classification,the definition and representation method of conceptual hierarchy,and criteria of evaluating classification scheme. According to the analysis,we briefly provede the architecture of data classifier. In the implementation of data classifier,we describe extraction and management of conceptual hierarchy for data,also design an automatic extraction algorithm for numeric data .In this section,we still provide the two algorithms of Concept-Based Attribute-Oriented Induction and evaluating classification scheme and the visualization of classification rule. Finally,the data classifier is tested in databas the results show that it is practical and its performance meet the requirement of designing.