Mining correlated high utility-frequent association rules under various discount notations

Association analysis is effective to explore relationships or similarities between items that are concealed in massive datasets. The uncovered associations can be characterized as association rules. i.e. discovering new-opportunities for cross-selling the product. Various algorithms elaborate high utility association rules as positive utility values. In real-life appliances, however, a high utility association rules may be associated with items having negative utility values when discounts are considered for certain products. This abundantly hinders their profits for various real-time appliances such as cross-selling or product recommendations so, finding high utility-frequent itemsets under various discount notations is significant for exploring interesting patterns. Also, a well-known constraint of association rules that are determined by using frequent patterns or utility patterns is that, they do not yield a measure of lift to find correlation between items. In this paper, we introduce a novel algorithm called HUFARM-N (High utility-frequent association rule mining with Negative utility values) which incorporates several expansions to mine high utility-frequent association rules that can meet the business profits ensuing to firms. Empirical analysis on real world datasets exhibits that, HUFARMN is highly capable and also enhances both execution time and memory usage.