Author:
Purnomo Jerry Dwi Trijoyo,Pratiwi Dea Restika Augustina
Abstract
Breast cancer is a malignant tumor that attacks breast tissue. This disease can be treated and managed properly if diagnosed at an early stage. An appropriate, fast and effective cancer stage detection algorithm is required so that patients can be treated precisely. In this study, the classification of breast cancer stages will be carried out using several machine learning methods. The number of patients in each stage is unequal or unbalanced as well. Therefore, the oversampling method with SMOTE is applied. The selection of the best parameters is done using 10-fold cross validation on the training data. Next, modeling was carried out using the Neural Network method, and K-Nearest Neighbor on training and testing data which had been oversampled with SMOTE. It was found that the neural network had a higher AUC value than k-Nearest Neighbor, namely 82.3% while k-NN was 80.8%.