Strategic Study of CAE >> 2002, Volume 4, Issue 5
A Study on Double Bases Cooperating Mechanism in KDD (Ⅱ)
Information and Engineering School, University of Science and Technology Beijing, Beijing 100083, China
Next Previous
Abstract
On the basis of paper ( I ) , this paper realized the reduction of the knowledge base and constructed the two coordinators and a component in the double-bases cooperating mechanism. Furthermore, the KDD* , which is the combination of KDD and double-bases cooperating mechanism, is constructed as an open and optimized enlargement of KDD. In the same time, this paper proposes a new association rule mining algorithm——Maradbcm which is based on the study on the inner mechanism, i. e. double base cooperating mechanism, of KDD and explains the algorithm of Maradbcm and Apriori respectively. Through the analysis of the algorithm and examples, it is shown that the main disadvantage of Apriori, overall database searching and deleting meaningful rules by minimum support threshold during creating large item set, can be overcome by Maradbcm algorithm. Then this paper proposed a new clustering method based on double-base cooperating mechanism. It is shown that the study of inner mechanism is important to the KDD mainstream development and it also provides a new path for the research in the area of knowledge discovery system.
Keywords
nowledge discovery ; double-base cooperating mechanism ; heuristic coordinator ; interruptive coordinator ; Maradbcm algorithm ; numerical domain division algorithm
References
[ 1 ] 杨炳儒, 王建新KDD中双库协同机制的研究 (Ⅰ) [J]中国工程科学, 2002, 4 (1) :41~51 link1
[ 2 ] AgrawalR , SrikantR .Fastalgorithmsforminingasso ciationrules[A].Proc20thInt’lConfVeryLargeDatabase[C], Santiago, Chile, 1994.487~499
[ 3 ] SavasereA , OmiecianskiE , navatheS .Anefficiental gorithmforminingassociationrules[A].ProcInt’1ConfVeryLargerDatabases[C], Zurich, 1995.423~444
[ 4 ] Yang, Bingru.KD (D&K) anddouble basescooperatingmechanism[J].JournalofSystemEngineeringandElectronics, 1999, 10 (2) :48~54
[ 5 ] BrinS , MotwaniR , SilversteinC .Beyondmarketbas kets:Generalizingassociationrulestocorrelations[A].ProcoftheACMSIGMOD [C], Montreal, Canada, 1996.255~276