Font Size: a A A

The Introduction Of The Constraint Factor, Bar Association Rule Mining Algorithm

Posted on:2006-03-11Degree:MasterType:Thesis
Country:ChinaCandidate:W Y ZhangFull Text:PDF
GTID:2208360182956384Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
Association rules mining must resolve the problem of efficiency. Due to that the data sets faced the task of data mining is large database or data warehouse which usually is composed of millions of records, so how to improve the efficiency and scalability of mining algorithm to low the complexity of calculation and improve the execute speed of the algorithm become the core problem in the research of association rules mining.At the same time, a problem existed in association rules mining is that those rules have high support usually isn't interested.Because the most of these redun- dant rules can be eliminated through the knowledge user had. Besides, user's requirement becomes more accurate. On this condition, we can constrain the rules. In this paper, I introduce the convinced factor and Min-interest, for, Minjnterest are often decided by specialist or user's knowledge, and convinced factor can solve the problerm which rule Bâ†'A and Aâ†'B what is a the really interesting rone. Compare BAR algorithm with Apriori algorithm, BAR algorithm improved much in the execute time. This paper introduce the association rules mining containing factor_constrains ,and discuss the problem that how to import factor constrain to BAR algorithm, and work out the BAR + algorithm, then compare the BAR+ algorithm with BAR algorithm, which testified the BAR+ algorithm can greatly improve the execute efficient of mining algorithm.
Keywords/Search Tags:Datamining, Bitmap Gruannular, BAR algorithm factor_constrain
PDF Full Text Request
Related items