A hybrid machine learning approach for enhanced skin cancer diagnosis using convolutional neural networks, support vector machines, and gradient boosting

This study investigates the effectiveness of a hybrid machine learning model for skin cancer diagnosis, integrating Convolutional Neural Networks, Support Vector Machines, and Gradient Boosting algorithms. By combining the strengths of each technique, the model seeks to improve diagnostic accuracy and reliability in clinical settings, addressing the challenges posed by traditional diagnostic methods. Utilizing the "Skin Cancer: Malignant vs. Benign" dataset, the hybrid model achieved an accuracy of 84%, with precision, recall, F1 score, and specificity recorded at 85%, 84%, 84%, and 83%, respectively. These results underscore the model’s potential to surpass single-algorithm approaches in detecting skin cancer, making it a promising tool for early diagnosis and better-informed clinical decision-making. The findings highlight the broader impact of advanced machine learning techniques in healthcare, particularly in oncology, by demonstrating how the integration of multiple algorithms can provide more accurate, scalable, and reliable diagnostic solutions. This research opens avenues for further exploration of hybrid models as a means to advance AI-driven diagnostic technologies in medical fields, with potential applications across various types of cancer detection. The source code for this study is available through a public GitHub repository, fostering transparency and further innovation in the field.