EARLY BREAST CANCER PREDICTION USING MACHINE LEARNING FRAMEWORK

Abstract

Breast cancer is considered as the second most diagnosed cancer in women caused by various factors like social, economic, or psychological. Although, often relies on mammography and biopsy reports, Machine learning and Artificial Intelligence have played a significant role in breast cancer research. This study highlights the importance of Machie learning in breast cancer prediction and proposes a Machine learning framework that comprises three models based on one of the popular Machine Learning algorithms - Support vector machine (SVM). An empirical study is performed using ‘linear’, ‘poly’ and ‘rbf’ kernels to evaluate the performance of Support Vector Machine in Model 1. To enhance the prediction performance, feature engineering techniques are employed to reduce model’s complexity as well as computation time in Model 2. As a base learner, Support Vector Machine is attempted to improve its prediction performance using ensemble learning followed by 10-fold cross-validation in Model 3. Results show that the Model 3 has greatly improved its accuracy up to 99.34% compared to Model 1 and Model 2 while SVM with ‘poly’ kernel is used as one of the diverse models of stacking, a highly accurate and balanced ensemble model.