Feature Level Ensemble Learning Technique for Cervical Cancer Cell Classification

The early detection of abnormal cells with Pap smear screening is a commonly used strategy for preventing cervical cancer, which is a major public health concern. In this study, we use the SIPaKMeD dataset to present an ensemble learning and multi - model ML learning strategy based on DL and ML for the categorization of cancerous cells in cervical cancer. Here we have use feature-level ensemble learning. ResNet, DenseNet, EfficientNet, and MobileNet were four pre-trained deep learning models that we utilised to extract features from the cervical cell pictures. These features were concatenated to create a feature vector, which was then classified using five base classifiers: Support Vector Machines, K-Nearest Neighbours (KNN), Random Forest (RF), Decision Trees (DT), and Logistic Regression (LR). The combined features were then applied to each classifier individually. The highest accuracy of 95% was achieved by the LR. The accuracy of the individual basic classifiers was 94,.4% (SVM), 87.2% (KNN), 76.5% (RF), and 74.6% (DT), respectively. Our suggested method has the potential to increase the precision of aberrant cell classification in Pap smear-based cervical cancer screening