Discovery of interesting effectiveness itemsets with several minimal thresholds using average constraint

Abstract

Discovery of interesting effectiveness itemsets has been a hot topics in data mining and knowledge discovery. Most previous researches related to effectiveness mining in literature employ a individual minimal threshold of effectiveness to decide on if an item is a interesting effectiveness item. Nevertheless, a individual minimal threshold of effectiveness could not express the varied natures of diverse items. In this paper, the author put forward a algorithm to discover interesting effectiveness itemsets with several minimal thresholds using average constraint. The reports of tests demonstrated that our algorithm is better than other baseline algorithms in performance.