Research On The Technology Of Association Rules In Data Mining

Association rules mining, as one of the most important contents in data mining, reveals the corelations between itemsets and can be widely applied to many fields such as market basket analysis, corelation analysis, classification, web-customised service, etc. Association rules mining was firstly proposed by R.Agrawal in 1993, in the last years the data mining research and the application domain active front. The typical Association rules mining algorithm is by R.Agrawal and so on Apriori Algorithm, its core technology for other each kind of Association rules mining excavation algorithm widespread use.However, with the distributed system being exist widely, using the traditional centralized association rule mining to discover useful patterns in distributed information system is not always feasible. Distributed association rules mining has thus emerged as an active subarea of association rules mining research. Based on the analysis of the existing distributed algorithms for mining association rules, an efficient distributed association rules mining algorithm(ED-ARM) is presented to fast find the large itemsets over the distributed transaction database system. The performance study and the test results show that ED-ARM algorithm is efficient and feasible.In addition, association rules updating, as another important content of the study of association rules mining, focuses on how to efficiently update the frequent intemsets on the changing of database or the mining parameters. The problem of Incremental updating of frequent itemsets is introduced by the numbers and corresponding algorithm FIUP (Frequent Itemsets Incremental Updating) is presented to handle the changing of the minimum support together with the transaction database including inserting and deleting the transactions in the database. The algorithm makes full use of the previous mining result to cut down the cost of updating frequent itemsets. At the end of the paper, some analyses and tests to the algorithm are offered, which shows that the algorithm is efficient and feasible.