Efficient Incremental High Utility Itemset Mining

High-utility itemset mining (HUIM) in transaction databases is an important data mining task with wide applications. However, most HUIM algorithms assume the unrealistic as-sumption that databases are static. To address this issue, algorithms have been designed to maintain high-utility item-sets in dynamic databases. However, these incremental al-gorithms still remain very costly in terms of execution time. In this paper, we address this problem by proposing an algo-rithm named EIHI (Efficient Incremental High-utility Item-set miner), which introduces several ideas to more efficiently maintain high-utility itemsets in dynamic databases. An experimental study on four datasets shows that EIHI is up to two orders of magnitude faster than the state-of-the-art HUI-LIST-INS algorithm