This study focused on post-menopausal women in the UKB cohort to explore the potential of using Machine Learning (ML) methods for feature selection to complement classical statistical methods for risk prediction of breast cancer. ML methods were used for feature selection, followed by classical Cox models for risk prediction. SHapley Additive exPlanation (SHAP) feature dependence plots were used to explore potential interactions between PRS and phenotypic features. Statistical considerations were provided before constructing classical Cox models to further investigate the potentially novel features selected by ML methods.