Ensemble Learning for Breast Cancer Detection: A High-Accuracy Model Integrating Clinical Insights

Breast cancer is one of the major global health issues, and early detection is the key to good patient outcomes. While interest in the application of various machine learning approaches to breast cancer detection is rife, more accurate and interpretable models are required to extract the potential added value that can be gained from using structured clinical data. This study aims to develop an advanced ensemble methodology for breast cancer detection using machine learning techniques, combining traditional statistical methods with modern ensemble learning. We employed an ensemble approach integrating Logistic Regression, Decision Tree, and Random Forest models, trained and validated on the Wisconsin Breast Cancer dataset. The proposed model achieved remarkable performance with $\mathbf{9 8. 8 2 \%}$ accuracy and $\mathbf{0. 9 9 4}$ AUC-ROC, outperforming several state-of-the-art methods. Feature importance analysis revealed mean concave points and worst radius as the most significant predictors, aligning with clinical understanding of breast cancer characteristics. The model’s high sensitivity (98.1%) and specificity ($\mathbf{9 6. 5 \%}$) demonstrate its potential for clinical application in early cancer detection, potentially enhancing diagnostic accuracy and contributing to improved patient outcomes in breast cancer management.