کنگره بین المللی پیشگیری از سرطان

صفحه اصلی / اولین کنگره بین المللی پیشگیری از سرطان

Ensemble-Based Breast Cancer Prediction Using Risk Factor Data

نویسندگان :

Hannah Yousefpour¹ Sekineh Asadi Amiri²

1- دانشگاه مازندران 2- دانشگاه مازندران

کلمات کلیدی :

Breast Cancer

چکیده :

Breast cancer, being one of the most commonly diagnosed cancers, poses a significant health challenge globally. It accounts for approximately 2.3 million new cases each year, as per the GLOBOCAN 2020 data. This study leverages machine learning to enhance early detection and prediction of breast cancer, utilizing a comprehensive dataset of risk factors. Normalization was applied on the dataset and then 80% of the data was used for training five ensemble-based algorithms: Random Forest, Rotation Forest, LightGBM, XGBoost, and CatBoost. These algorithms were assessed using four performance metrics: accuracy, precision, recall, and F1-score. CatBoost emerged as the leading algorithm, achieving an impressive accuracy of 87.17%, precision of 86.25%, recall of 87.17%, and F1-score of 86.37%. These results highlight CatBoost's superior performance in accurately detecting breast cancer cases, crucial for effective treatment and patient management. By integrating advanced machine learning methodologies, this research contributes significantly to improving early detection and diagnostic accuracy, thereby enhancing patient outcomes and supporting better clinical decision-making in the management of breast cancer.