wFDT-Weighted Fuzzy Decision Trees for Breast Cancer Survivability Analysis

Abstract

Accurate and less invasive personalized predictive medicine can spare many breast cancer patients from receiving complex surgical biopsies, unnecessary adjuvant treatments and its expensive medical cost. Cancer prognosis estimates recurrence of disease and predict survival of patient; hence resulting in improved patient management. To develop such knowledge based prognostic system, this thesis examines potential hybridization of accuracy and interpretability in the form of Fuzzy Logic and Decision Trees, respectively. Effect of rule weights on fuzzy decision trees is investigated to be an alternative to membership function modifications for performance optimization. Experiments were performed using different combinations of: number of decision tree rules, types of fuzzy membership functions and inference techniques for breast cancer survival analysis. SEER breast cancer data set (1973-2003), the most comprehensible source of information on cancer incidence in United States, is considered. Performance comparisons suggest that predictions of weighted fuzzy decision trees (wFDT) are more accurate and balanced, than independently applied crisp decision tree classifiers; moreover it has a potential to adapt for significant performance enhancement.