Skin Cancer Prediction Using Machine Learning Algorithms

Skin cancer is anomalous development of skin cells which are prevalent in human beings. It can occur in both cases, when the skin is exposed or when not exposed to sunlight. Due to the exposure of pollutants, chemicals and cosmetics, the condition of skin is becoming worse. The medical field is improving with the innovation of very new technologies. The early diagnosing is a vital role in determining the diseases and getting cured, but even highly experienced doctor could not identify and classify the skin diseases. Hence, the computer aided melanoma detection is essential for early identification of disease. Through this approach, the mortality of cancer patient can be reduced. The proposed architecture is focused on the automated system for the prediction of dermatological diseases. This model is designed into five phases compromising of data collection, preprocessing, segmentation, feature extraction and prediction. The outcome of CNN, AdaBoosting, Gradient Boosting and Decision Tree algorithms were compared in this study. The proposed machine learning model requires less intervention of human in the cancer prediction process.