Performance evaluation of machine learning techniques for breast cancer detection using WDBC dataset
Indu Chhillar, Ajmer Singh · AIP conference proceedings · 2024
Recently, the study of machine learning has gained prominence in the medical field.One of the most challenging tasks in machine learning is creating accurate and reliable classifiers for medical applications.To find the best classifier, a comparative analysis of the most prominent machine learning classifiers-SVM, DT, RF, NB, KNN, XGBoost, AdaBoost, and CatBoost-has been done using the Wisconsin Diagnostic Breast Cancer (WDBC) dataset.The performance of classifiers is evaluated in terms of accuracy, precision, recall, AUC, confusion matrices, and ROC.The results of the analytical comparison revealed that the SVM and RF classifiers outperformed other classifiers.Furthermore, the WDBC dataset used in this study has a class imbalance issue (malignant 212; benign 357).Imbalanced data sets produce skewed outcomes and deceptive accuracy.Data resampling was performed to fix the problem of class imbalance.