Abstract:

Association rule is one of the most important patterns for the research of data mining, and there are now mature systems and approaches for mining the  Boolean association rules while there are the opposite conditions for mining quantitative association rules. This paper presents a new algorithm QARMM  which makes use of matrix to store data and transforms the process of generating frequent item sets to the process of calculating matrix vectors. The QA RMM algorithm overcomes the drawbacks of the SL IG algorithm, needs to run only once to mine all the association rules,and has better efficiency than th e Apriori algorithm on equivalent datasets.

Key words: quantitative association rule, recognizable vector, MFS