Association Rule Mining Based HotSpot Analysis on SEER Lung Cancer Data

Abstract

The authors analyze the lung cancer data available from the SEER program with the aim of identifying hotspots using association rule mining techniques. A subset of 13 patient attributes from the SEER data were recently linked with the survival outcome using prediction models, which is used in this study for segmentation. The goal here is to identify characteristics of patient segments where average survival is significantly higher/lower than average survival across the entire dataset. Automated association rule mining techniques resulted in hundreds of rules, from which many redundant rules were manually removed based on domain knowledge. Further, association rule mining based hotspot analysis was also conducted for conditional survival patient data, i.e., in cases where patients have already survived for a year after diagnosis. The resulting rules conform with existing biomedical knowledge and provide interesting insights into lung cancer survival.