Association Mining for Super Market Sales using UP Growth and Top-K Algorithm

Abstract

Frequent itemsets(HUIs) mining is an evolving field in data mining, that centers around finding itemsets having a utility that meets a user-specified minimum utility by finding all the itemsets. A problem arises in setting up minimum utility exactly which causes difficulties for users. By setting minimum utility underneath average, too many incessant itemsets will be generated, which in turn will make the mining process quite inefficient. No frequent itemsets will be found if the minimum utility is set too huge. The research focuses on generating frequent itemsets by using the transaction weighted utility of each product. While using UP growth methodology for discovering high utility items from large datasets it takes more time and consumes more memory due to which it is less efficient. So to overcome these drawbacks of UP growth we use the Top-K algorithm which makes it more scalable and efficient. Therefore, we use the Top-K algorithm which does not require a minimum threshold.