HMM-fuzzy model for breast cancer diagnosis
Md. Rafiul Hassan, Rezaul K. Begg, Yosry Morsi, Kate Lynch · Swinburne figshare (Swinburne University of Technology) · 2006
This paper presents a Hidden Markov Model (HMM) based fuzzy model for breast cancer recognition and classification. The feasibility of using HMM for generating a minimum number of fuzzy rules for a given problem has been introduced in a recent study. In this study, we apply this new approach to investigate whether or not it can differentiate effectively the two important breast cancer types: benign and malignant. Artificial neural networks (ANNs) have been proposed to improve the diagnosis of breast cancers. Fuzzy logics have the advantage compared to the ANNs in that it (Fuzzy model) is interpretable and can represent the solution of a problem with reduced computational complexity. In this paper, we demonstrate the suitability on the use of fuzzy model over ANNs for breast cancer diagnosis, as the model was able to provide a comparable classification performance (accuracy) with reduced complexity.