Ensemble of Random Forest and Adaboost Algorithms for Human Skin Disease Classification

Skin diseases are a common health issue that is often overlooked, despite the fact that they can develop into serious conditions if not addressed early. To assist in early detection, this study developed a skin disease classification system based on digital images using an ensemble learning approach with Random Forest and AdaBoost algorithms. The dataset used was sourced from Kaggle and includes 9 types of skin diseases: Actinic Keratosis, Atopic Dermatitis, Benign Keratosis, Dermatofibroma, Melanocytic Nevus, Melanoma, Squamous Cell Carcinoma, Tinea Ringworm Candidiasis, and Vascular Lesion. The system was developed as a website that allows users to upload images for automatic classification. Test results show that the Random Forest algorithm achieved an accuracy of 65.66%, precision of 67.93%, recall of 65.66%, and a misclassification error of 34.34%. Meanwhile, the combination of Random Forest and AdaBoost improved the accuracy to 71.69%, precision to 75.23%, recall to 75.23%, and reduced the misclassification error to 28.31%. The 6.03% increase in accuracy indicates that the ensemble approach provides better and more stable classification compared to single algorithms. This research is expected to support the development of early detection systems for skin diseases based on artificial intelligence.