Predicting of Melanoma Skin Cancer Using Machine Learning Methods.

Cancer disease is known as the second highest cause of death in the world. One of the most common types of this disease is skin cancer. As in almost all cancer diseases, early diagnosis and treatment of skin cancer is of great importance. In the process of diagnosing cancer, Machine Learning-based methods are also widely used in addition to traditional methods. The most important advantage of these methods is that they eliminate or minimize human errors that may arise during the cancer diagnosis process. In this study, skin cancer was diagnosed with K-Nearest Neighbor (KNN), Naive Bayes (NB), Random Forest (RO), Logistic Regression (LR), and Artificial Neural Networks (ANN) methods using images taken from patients. In these algorithms, the training and testing process was run, the results were analyzed and models were created. With these models, benign and malignant lesions were compared and skin cancerous lesions were detected and success percentages were revealed. As a result, the best results were obtained using the ANN method, with a training percentage of 99.1% and a testing percentage of 98.6%. When different inputs were given to the created and proposed ANN model, it was observed that the model predicted skin cancer with a high accuracy rate. The results obtained revealed the success of the study and that machine learning methods are a usable method in the cancer diagnosis process. [ABSTRACT FROM AUTHOR]