STUDY OF HIGH UTILITY PATTERN MINING ALGORITHMS AND COMPARISION OF D2HUP AND MAHUSP ALGORITHMS

Abstract

Discovering interesting patterns and useful knowledge from massive data has become an important data mining task. These days, we come across a lot of things that have profit technically referred as external utility, value greater than the other item sets in the database. Utility mining is an important topic in data mining and has received extensive research in last few years. In utility mining, each item is associated with a utility that could be profit, quantity, cost or other user preferences. Objective of Utility Mining is to identify the item sets with highest utilities. High utility itemset mining is an extension to the problem of frequent pattern mining. Many algorithms have been proposed in this field in the recent years. In this paper we emphasis on an emerging area called High Utility Mining which not only considers the frequency of the itemsets but also considers the utility associated with the itemsets. In High Utility Itemset Mining the target is to identify itemsets that have utility value greater than the threshold utility value. In this paper we present a review of the various techniques and current scenario of research in mining high utility itemset also presented advantages and limitations of various techniques for High Utility Itemset Mining. We mainly focus on the D2HUP and MAHUSP approache and algorithms for high utility pattern mining with less memory utilization.