Improving breast cancer classification with a novel VGG19-based ensemble learning approach

Abstract

Breast cancer is one of the most life-threatening diseases, particularly affecting women, highlighting the importance of early detection for improving survival rates. In this study, we propose a novel diagnostic framework that combines a modified VGG19 architecture with Bagging ensemble learning, using three base classifiers: decision tree (DT), logistic regression (LR), and support vector machine (SVM). We also compare this approach with twenty-four hybrid models, integrating various convolutional neural network (CNN) architectures (ResNet50, VGG19, ConvNextBase, DenseNet121, EfficientNetV2B0, EfficientNetB0, MobileNet, and NasNetMobile) with Bagging ensemble methods. Our results show that the proposed model outperforms all other architectures, especially when combined with SVM, achieving accuracy of 97% on the fine needle aspiration cytology (FNAC) dataset and 90% on the International Conference on Image Analysis and Recognition (ICIAR) dataset. This framework demonstrates strong potential for improving early breast cancer diagnosis.