Developing High Risk Clusters for Chronic Disease Events with Classification Association Rule Mining

Song, S., Warren, J. and Riddle, P.

Association Rule Mining (ARM) is a promising method to provide insights for better management of chronic diseases. However, ARM tends to give an overwhelming number of rules, leading to the longstanding problem of identifying the �interesting� Rules for knowledge discovery. Therefore, this paper proposes a hybrid clustering-ARM approach to gain insight into a population�s pattern of risk for a chronic disease related adverse event. Our current experiment is based on the Framingham Heart Study dataset and we focus on Myocardial Infarction (MI, �heart attack.�) as the adverse event. Association rules indicative of MI are developed from training data and clustered based on commonality of cases satisfying the rule antecedents. Test cases are then assigned to the rule clusters to provide sets of at-risk patients sharing common MI risk factors. We demonstrate this approach for a range of clustering methods and cluster counts, illustrating some of the derived participant sets.1

Cite as: Song, S., Warren, J. and Riddle, P. (2014). Developing High Risk Clusters for Chronic Disease Events with Classification Association Rule Mining. In Proc. Seventh Australasian Workshop on Health Informatics and Knowledge Management (HIKM 2014) Auckland, New Zealand. CRPIT, 153. Warren, J. and Gray, K. Eds., ACS. 69-78

(from crpit.com) (local if available)