Abstract

Breast cancer remains one of the leading causes of mortality among women globally, emphasizing the need for early and precise diagnostic systems. Traditional machine learning models, while effective, often function as black boxes, offering limited interpretability to healthcare professionals. Despite advancements in diagnostic tools, there remains a gap in delivering models that are both highly accurate and explainable. Existing models tend to prioritize predictive performance over transparency, making it difficult for clinicians to trust and adopt them in real-world scenarios. This work proposes a Fuzzy Rule-Based Modelling (FRBM) approach for breast cancer prediction that balances accuracy with interpretability. The proposed system translates numerical input data into linguistic fuzzy sets and derives inference rules using a Sugeno-type fuzzy inference system. Feature selection is carried out using a combination of correlation-based methods and expert knowledge to ensure only relevant diagnostic attributes are used. The model generates understandable IF-THEN rules, providing clinicians with clear decision logic. The dataset used is the Wisconsin Diagnostic Breast Cancer (WDBC) dataset from the UCI repository. The proposed fuzzy model achieved an accuracy of 97.6%, outperforming traditional models such as Support Vector Machines (SVM) and Decision Trees (DT), which achieved 94.8% and 93.5%, respectively. Additionally, the fuzzy system demonstrated a high F1- score of 0.96 and excellent interpretability, enabling users to understand and validate predictions.

Authors

S. Savitha¹, V. Umadevi²
K.S.R. College of Engineering, India¹, Arunai Engineering College, India²

Keywords

Breast Cancer Prediction, Fuzzy Rule-Based System, Interpretability, Explainable AI, Medical Diagnosis