Literature Review on Efficient Algorithms for Mining High Utility Itemsets from Transactional Databases

Abstract

This paper presenting a survey on finding itemsets with high utility. For finding itemsets there are many algorithms but those algorithms having a problem of producing a large number of candidate itemsets for high utility itemsets which reduces mining performance in terms of execution. Here we mainly focus on two algorithms utility pattern growth (UP-Growth) and UP-Growth+. Those algorithms are used for mining high utility itemsets, where effective methods are used for pruning candidate itemsets. Mining high utility itemsets Keep in a special data structure called UP-Tree. This, compact tree structure, UP-Tree, is used for make possible the mining performance and avoid scanning original database repeatedly. In this for generation of candidate itemsets only two scans of database. Another proposed algorithms UP Growth+ reduces the number of candidates effectively. It also has better performance than other algorithms in terms of runtime, especially when databases contain huge amount of long transactions. Utility-based data mining is a new research area which is interested in all types of utility factors in data mining processes. In which utility factors are targeted at integrate utility considerations in both predictive and descriptive data mining tasks. High utility itemset mining is a research area of utility based descriptive data mining. Utility based data mining is used for finding itemsets that contribute most to the total utility in that database.