Enhanced Skin Cancer Classification Using Deep Learning and Gradient Boosting Techniques

Skin cancer is a common and potentially fatal disease that requires early and accurate diagnosis for effective treatment. This paper presents a machine learning-based approach for classifying skin cancer using dermoscopic images from the HAM10000 dataset. We utilize pre-trained models, specifically ResNet-50 and DenseNet-201, for feature extraction, followed by dimensionality reduction using Principal Component Analysis (PCA). The resulting feature set is then classified using CatBoost, a state-of-the-art gradient boosting algorithm. The proposed model achieved an impressive accuracy of 98.15% and a log loss of 0.2141. We compare our results with nine recent studies and demonstrate the superior performance of the proposed approach. This study highlights the potential of integrating advanced deep learning techniques and boosting algorithms to enhance the accuracy and reliability of automated skin cancer diagnosis systems.