Research on improvement of objective interestingness measures for association rules

Association rule is an important research topic in the fields of data mining, and objective interestingness is the measure to evaluate the quality of association rules. But at this stage, objective interestingness measures cannot identify the valid association rules in the datasets accurately. Some measures may lead to the explosion of rules´ number. In response to these problems, this paper introduces the related concepts of distance relevancy and entropy from the fields of statistics and information theory, and puts forward two new measures called Newrelevancy and NewI by improving the two measures. Newrelevancy is used to find frequent itemsets, and NewI is used to mine the strong association rules in these found frequent itemsets. Data analysis shows that compared to traditional measure framework, the new framework made up of Newrelevancy and NewI has a better evaluation effect.