Breast Cancer Relapse Prognosis by Classic and Modern Structures of Machine Learning Algorithms

According to medical reports, cancers are big problems in the world society. In this paper we are supposed to predict breast cancer recurrence by multi-layer perceptron with two different outputs, a deep neural network as a feature extraction and multi-layer perceptron as a classifier, rough neural network with two different outputs, and finally, support vector machine. Then, we compare the results achieved by each method. It can be understood that rough neural network with two outputs leads to the highest accuracy and the lowest variance among other structures.