Efficient Frequent Itemset Mining Using Global Profit Weighted (GPW) Support Threshold

Discovery of efficient association rules has been found useful in many applications. However, without fully considering the importance and significance of items and transactions, it is noted that some discovered rules may be expired from users' interest. This paper proposes a novel technique to solve this problem. Utility measures play an important role in data mining, regardless of the kind of patterns being mined. In this paper we present a new approach for finding frequent itemsets by using level of profit based support threshold instead for minsupport threshold as in classical association rule mining algorithm. The profit or the importance of the items in the itemsets is computed through efficient decision mechanism analytic hierarchy process. We introduce a new algorithm to find Global Profit Weight (GPW) measure for the item in the transaction which is based on the item characteristic by using statistical and analytical methodology.