Abstract
Traditional association rule mining algorithms often have difficulty handling questions that are implicitly related, producing rules that are very accurate but are so obvious as to be completely useless to researchers. This problem is compounded by the fact that standard objective measures for interestingness often capture the wrong information and are maximized by these obvious rules. We propose an enhancement to standard association rule mining that uses clustering to identify related questions to pre-prune rules involving similar questions, which are less likely to be subjectively interesting. This enhancement reduces the search space of rules, improving the algorithm's efficiency. In addition, the resulting rule list has a higher concentration of diverse rules, more likely to be useful to researchers. We demonstrate this improvement relative to existing algorithms on two real-world, public health questionnaires.