Ensemble Machine Learning Models for Breast Cancer Identification

The advances in the Machine Learning (ML) domain, from pattern recognition to computational statistical learning, have increased its utility for breast cancer as well by contributing to the screening strategy of diverse risk factors with complex relationships and personalized early prediction. In this work, we focused on Ensemble ML models after using the synthetic minority oversampling technique (SMOTE) with 10-fold cross-validation. Models were compared in terms of precision, accuracy, recall and area under the curve (AUC). After the experimental evaluation, the model that prevailed over the others was the Rotation Forest achieving accuracy, precision and recall equal to 82% and an AUC of 87.4%.