COMPARISON OF SUPPORT VECTOR MACHINE AND DECISION TREE METHODS IN THE CLASSIFICATION OF BREAST CANCER

Abstract

One of the most dangerous cancers in the world is breast cancer. This cancer occurs in many women, in some cases this cancer can also affect men, but it is very rare. The effects of this cancer are very dangerous for humans, in the worst case it can lead to death. So that serious prevention is needed against this cancer. One prevention can be done by early detection. This study aims to implement machine learning methods to detect breast cancer in women. The algorithms used are Support Vector Machine (SVM) and Decision Tree (DT). After classifying the data provided, a comparison is made to find out which machine learning method has the best performance. The data used comes from the Gynecology Department of the University Hospital Center of Coimbra (CHUC), and can be downloaded for free on the UCI repository website. The results of this study indicate that the SVM algorithm with feature selection obtains the best classification results by obtaining an accuracy of 87.5%, a sensitivity of 90%, and a specificity of 85%. Thus this research obtains good results to be able to help provide solutions to detect breast cancer.