Enhancing the pruning performance in on-shelf utility mining

In real-world applications, not all the products in stores are always on shelf for sale. On-shelf utility mining was thus proposed to deal with the problem. In this study, we extend our previous approaches with an on-shelf utility upper bound to early prune the unpromising candidates in the mining process. Especially, all candidates generated in each time period are utilized to get more accurate upper bound values of itemsets. Experimental results also show its performance.