Abstract
Breast cancer is one of the most prevalent types of cancer among women. With increased emphasis towards cancer related research, many data-driven research works have been conducted for classifying cancer diagnosis, survival, or recurrence. Unlike existing literature, this study aims to discover interesting subgroup patterns of long-term and short-term survival from the breast cancer incidence data of the SEER (Surveillance, Epidemiology, and End Results) Program. We present a rule induction method for subgroup discovery, which can effectively find subgroup patterns by focusing on local exceptionality detection in contrast to global models. The significance of subgroup patterns discovered is examined with statistical tests. Furthermore the characteristics of two exceptional high and low survival groups are compared by examining the descriptive statistics of prognostic factors in each group. The case study’s results show that the proposed subgroup mining and statistical test approach is a promising technique for clinical and medical data analytics.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call