Research Of Dynamic Association Rules

The rapid progress of data collection and data strorage technology allows organizationesaccumulate massive amounts of data. However, the extraction of useful information has becomea huge challenge. Data mining technology has provided us with a good way out of thispredicament, and association rules mining is one of the most active data mining researchmethods. The problem of mining association rules has been studied by many researchers and alot of research has been done, meanwhile a lot of new algorithms are made, but almost all of thealgorithms assume that the rules are valid all the time and don't change.In behavior analysis, the amount of data is very large and the support value of associationrules are very small. When evaluating association rules, confidence value is condidered. Sincethe data warehouse need to be updated frequently, the confidence of association rules that havebeen found change with the data waurehouse. In this paper, when describing the associationrules changes, the support count and confidence count vector pair was introduced forevaluation.The main work of this paper focuses on comparing the efficiency of traditional datawarehousing and consolidation with some of the features of the Oracle,and improving thetraditional evaluation method of association rules.Meanwhile when the data set has dynamicchanges,the method of mining the dynamic association rules is proposed.