Search

Mining high utility quantitative itemsets (HUQIs) is now a novel research topic in data mining field, which consists of discovering sets of items having a high utility (e.g. high profit) and providing information about quantities of items in each itemset. In market analysis, it could supply for decision-makers that shopping behavior could bring high profit to the company. For example, the customers purchase M to N units of a product A and purchase P to Q units of a product B at the same time. However, mining HUQIs using existing algorithms remains very computationally expensive and makes the results hard to be utilized by users. In view of this, we propose a novel algorithm named HUQI-Miner (High Utility Quantitative Itemsets Miner) for efficiently mining HUQIs in databases. Experimental results on both real and synthetic datasets show that HUQI-Miner outperforms the state-of-the-art algorithms in terms of both execution time and memory usage.

High utility quantitative itemset mining refers to discovering sets of items that carry not only high utilities (e.g., high profits) but also quantitative attributes. Although this topic is very important to many applications, it has not been deeply explored and existing algorithms for mining high utility quantitative itemsets remain computationally expensive. To address this problem, we propose a novel algorithm named VHUQI (Vertical mining of High Utility Quantitative Itemsets) for efficiently mining high utility quantitative itemsets in databases. VHUQI adopts a vertical representation to maintain the utility information of itemsets in databases with several effective strategies integrated to prune the search space. The experimental results on both real and synthetic datasets show that VHUQI outperforms the state-of-the-art algorithms substantially in terms of both execution time and memory consumption. Keywords—Quantitative itemset mining; high utility itemset mining; high utility quantitative itemset mining; utility mining