关联规则采掘的-计算机应用技术专业论文.docxVIP

些!!圣丝! 些!!圣丝! 塑鋈查兰竖主兰垡丝奎 ABSTRACT With the capacity of collecting and storing data improving，scholars from artificial intelligent，database and other related field are paying great attention to data mining Mining association rules is an important data mining problem Its main purpose is discovering relativity between items or attributes that stored in the databases．There have been many algorithms proposed for it．Based on summarizing and analyzing these mining algorithms，seyeral efficient algorithms for mining association rules arc presented．The main works of this paper are as following： The Apriori algorithm is improved．A more efficient algorithm，Apriorilpv，is proposed to overcome the weakness that the algorithm Apriori does not make full advantage of result of former scanning the database A more efficient mining algorithm，LIDBSO，which is based on set operation，is given．Contrasting to level—wise mining algorithm that need scan the database more times，it only need scan the database once．So its mining efficiency is increased much more． An efficient parallel algorithm，PMAR，which is based on share nothing architechure in multiple computers，is given to settle the problem of mining association rules in a distributed database or a very large database An incremental updating algorithm for mining association rules，EUIA，which is based on algorithm LIDBSO，is proposed．It can sovle the three kinds of incremental updating problem for mining association rules in central databases The prototype system，VSMAR，which is visualization System for Mining Association Rules，is designed Moreover,A hybrid algorithm，ALHybrid，which is made of the algorithm LIDBSO and the algorithm Apriorilpv,is designed．Two algorithm，MQC—Apriori and MQC～LIDBSO，are given to mining multiple level quantitative association rules with item constrains．Two algorithm，PMARⅡand PMARm，which are based on different se