Breast Cancer Classification using XGBoost

Abstract

Breast cancer continues to be one of the foremost illnesses that results in the deaths of numerous women each year. Among the female population, approximately 8% are diagnosed with Breast cancer (BC), following Lung Cancer. The alarming rise in fatality rates can be attributed to breast cancer being the second leading cause. Breast cancer manifests through genetic transformations, persistent pain, alterations in size, color (redness), and texture of the breast's skin. Pathologists rely on the classification of breast cancer to identify a specific and targeted prognosis, achieved through binary classification (normal/abnormal). Artificial intelligence (AI) has been employed to diagnose breast tumors swiftly and accurately at an early stage. This study employs the Extreme Gradient Boosting (XGBoost) machine learning technique for the detection and analysis of breast cancer. XGBoost provides an accuracy of 94.74% and recall of 95.24% on Wisconsin breast cancer Wisconsin (diagnostic) dataset.