Mining High Transaction-Weighted Utility Itemsets

In this paper, we design a new kind of patterns, named high transaction-weighted utility itemsets, which considers not only individual profits and quantities of the items in a transaction, but also the contribution of each transaction in a database. We also propose a two-phased mining algorithm to discover high transaction-weighted utility itemsets. The experimental results on synthetic datasets show the proposed approach has a good performance.