Data mining aims to discover interesting and useful knowledge out of great amount of data so that decision making can be done easily. Among important research methods, Mining Association Rule can be utilized to describe association between different valuable data. Traditional ARM algorithm known as Apriori uses single minimum support which is found not applicable since items are different in nature and each item must be given importance, thus, resulted to “rare item and rules explosion” dilemma. In this paper, new method in extracting frequent itemset has been proposed. Every item is given a minimum support value in order to mine items with high and low supports. The proposed approach utilized synthetic and real datasets and was evaluated in terms of the generated rules and processing time. Experimental results show that the proposed approach improves the performance of Apriori in discovering association rules and processing time.
Volume 12 | 01-Special Issue