A Conceptual Approach to Temporal Weighted Itemset Utility Mining

Conventional Frequent pattern mining discovers patterns in transaction databases based only on the relative frequency of occurrence of items without considering their utility. Until recently, rarity has not received much attention in the context of data mining. For many real world applications, however, utility of itemsets based on cost, profit or revenue is of importance. Most Association Rule Mining (ARM) algorithms concentrate on mining frequent itemsets from crisp data and recently, use of discrete utility values. Unfortunately, in most real-life applications, use of discrete valued utilities alone is inadequate. In many cases where these values are uncertain, a fuzzy representation may be more appropriate. An interesting extension to ARM is including the temporal dimension. Traditional ARM does not use time; however, the real application data always changes with time. Discovering temporal association rules that hold in given time intervals may lead to more useful information. As real-world problems become more complex, temporal rare itemset utility problems become inevitable to solve. To handle uncertainty, temporal itemset utility mining with fuzzy modeling allows item utility values to assume fuzzy values and be dynamic over time. In this paper, we present a theoretical conceptual approach to Temporal Weighted Itemset Utility Mining.