Fuzzy Expert System for Prognosis of Breast Cancer Recurrence

Abstract

Data mining techniques along with fuzzy logic, play an important role in decision-making applications with imprecise and uncertain knowledge. A fuzzy expert system models uncertain knowledge as a set of fuzzy rules and performs reasoning more accurately. This paper presents a fuzzy expert system for breast cancer prognosis, which is capable enough to capture the inherent ambiguity and imprecision of the breast cancer data. For this purpose we used UCI Machine Learning Repository, Breast Cancer Dataset, and proposed a new method of data mining which is a combination of decision tree and association rule mining. Through this new method knowledge acquisition was performed. Using fuzzy approximate reasoning, we achieved an accuracy rate about 93% for breast cancer recurrence event, which in comparison with other data mining methods can be considered as a remarkable progress.