Association Rule Mining (ARM) is a well-studied technique that identifies frequent itemsets from datasets and generates association rules by assuming that all items have the same significance and frequency of occurrence without considering their utility. But in a number of real-world applications such as retail marketing, medical diagnosis, client segmentation etc., utility of itemsets is based on cost, profit or revenue. Utility Mining aims to identify itemsets with highest utilities by considering profit, quantity, cost or other user preferences.

Rare items are items that occur less frequently in a transaction dataset. High Utility Itemsets may either be frequent or rare. Similarly rare itemset may be of high or low utility.

In many real-life applications, high-utility itemsets consist of rare items. Rare itemsets provide useful information in different decision-making domains, customers purchase microwave ovens or plasma televisions rarely as compared to bread, washing powder, soap etc. The former may yield more profit for the supermarket than the latter. Koh and Rountree (2005) proposed a modified apriori inverse algorithm to generate rare itemsets of user interest.

In this paper, the authors propose a High Utility Rare Itemset Mining [HURI] algorithm that uses the concept of apriori inverse, for generating high utility rare itemsets of users’ interest[Koh and Rountree (2005)]. We demonstrate the approach with a synthetic dataset. Apriori inverse is used to find only the rare itemsets. HURI is used to find those rare itemsets, which are of high utility according to users’ preferences, i.e., algorithm for generation of rare itemsets is extended to find high-utility rare itemsets.