Classification of Skin Diseases Using Ensemble Method

Abstract

Skin diseases are serious universal health issue which is associated with large group of human beings. In recent years, the fast evolution of technologies in application of different data mining techniques has improved their accuracy more and more in classification of dermatological predictions. The development of machine learning techniques has an immense importance so that can effectively differentiate the classification of skin diseases. In this paper, we develop a new method using three data mining techniques and then establish an ensemble approach. This ensemble approach uses voting consisting of all three different data mining algorithms like random forest, decision tree and support vector machine as a single unit. By using various data mining techniques, dermatology data is analyzed to diagnose skin diseases. Based on machine learning, the proposed ensemble approach was evaluated on dermatology dataset and categorized the type of skin diseases into seven separate groups, including Pityriasis rubra, Lichen planus, Rosea pityriasis, Healthy skin, Psoriasis, Chronic dermatitis, Seborrheic dermatitis.