Skin Cancer Diagnosis Using an Improved Ensemble Machine Learning model

In recent years skin cancer is becoming more and more threatening because of its fast and significant spread worldwide. This evidence has increased interest and efforts in the development of automatic diagnostic computational systems to assist early diagnosis. Several approaches have been proposed to assist in skin lesion diagnosis which used machine learning and ensemble learning. In some cases, a classifier can correctly predict the output class while others fail and vice versa. So the idea is to use different machine learning and ensemble learning to classify skin cancer. In this paper, we propose an improved ensemble learning method to classify skin cancer. Features used are the best combination of extracted features from different characteristics, i.e., shape, color, texture, and skeleton of the lesion, then we classify these features using different algorithms to predict the classes. Globally, the experimented results show a promoting result.