A share strategy for utility frequent patterns mining

Frequent pattern mining and utility mining have been studied popularly. However, frequent pattern mining only mines frequent patterns without considering the different utility values of individual items and utility mining focuses on identifying the patterns with high utilities but no guarantee their frequencies. In this paper, we introduce a utility frequent pattern mining model based on a share strategy to find the combination of items with high frequencies and utilities. This model first find all patterns with a given minimum support threshold. In this step, a share strategy gives a way to share most of the results from the previous mining process instead of separating them distinctively, thereby dramatically reducing the cost of computation. And then all patterns that do not satisfy a user specified utility are pruned. The performance study shows that the share strategy is efficient for utility frequent patterns mining.