Meme Kanseri Tanısı İçin Özniteliklerin Öneminin Değerlendirilmesi Üzerine Bir Çalışma

Abstract

Breast cancer, one of the most common types of cancer, is a deadly disease affecting women. The importance of attributes was investigated by using the Recursive Feature Selection based on feature selection on Wisconsin breast cancer dataset, and then the machine learnings were performed by utilizing Random Forest and Logistic Regression classifier algorithms in the proposed study. The learning process involving training and testing phases was performed by utilizing the 5-fold cross-validation technique. Experimental studies showed that the best classification performance (98% accuracy) was achieved by applying the Random Forest algorithm.