A comparative performance of breast cancer classification using hyper-parameterized machine learning models

Breast cancer is the second most common cancer and has the second-highest mortality rate in women among all cancer types. Accurate cancer diagnosis plays a great part in breast cancer treatment. The application of machine learning methods in cancer classification has grown popular and has provided an accurate classification of malignant (cancerous) and benign (non-cancerous) breast cancer. This paper presents the application of three machine learning methods to classify malignant and benign breast cancer. The three machine learning methods used in this study are Support Sector Machine (SVM), Logistic Regression (LR), and Neural Network (NN) for breast cancer classification. For each machine learning method, multiple models had been tested with every model having a unique set of parameter values. This study used the breast cancer Wisconsin diagnostic (BCWD) dataset. The performance of the models is evaluated using the k-fold cross-validation technique and confusion matrix. The result shows that SVM outperformed both LR and NN in terms of classification accuracy, precision, recall, and specificity with k-fold cross validation technique. On the other hand, when the train-test split was used to validate the proposed model, the NN outperformed both SVM and LR achieving accuracy of 99.4%.