Interestingness of association rules using symmetrical tau and logistic regression

Abstract

While association rule mining is one of the most popular data mining techniques, it usually results in many rules, some of which are not considered as interesting or significant for the application at hand. In this paper, we conduct a systematic approach to ascertain the discovered rules and provide a rigorous statistical approach supporting this framework. The strategy proposed combines data mining and statistical measurement techniques, including redundancy analysis, sampling and multivariate statistical analysis, to discard the non significant rules. A real world dataset is used to demonstrate how the proposed unified framework can discard many of the redundant or non significant rules and still preserve high accuracy of the rule set as a whole.