<scp>DHHoE</scp>: Deep hybrid homogenous ensemble for digital histological breast cancer classification

The progress of deep learning architectures, machine models and pathology slide digitization is an encouraging step toward meeting the growing demand for more precise classification prediction diagnosis breast tumours. BreakHis dataset with four magnification factors (40X, 100X, 200X 400X), as well seven architectures used feature extraction (DenseNet 201, Inception ResNet V2, V3, 50, MobileNet V2,VGG16 VGG19), (MLP, SVM, DT, KNN), two combination rules (hard weighted voting) were investigated in this paper to design evaluate a new proposed approach consisting building hybrid homogenous ensemble. Additionally, best compared stacked, bagging, boosting, heterogenous ensemble choose strategy techniques. performance measures accuracy, precision, recall, F1-score empirical evaluations, 5-fold cross-validation, Scott Knott statistical test, Borda Count voting method. results demonstrated approach's potential since it outscored both singles other strategies, achieving accuracy values 98.3% 97.7% MFs 40X, 100X 200X, 400X, respectively. that ensembles are impactful histopathological cancer images classification, they provided promising tool assist pathologists cancer.