<![CDATA[Breast cancer is one of the diseases with the highest incidence and mortality rates among women, requiring methods that can support fast and accurate detection. This study aims to compare the performance of the Decision Tree C4.5 and Neural Network algorithms in breast cancer classification using the Breast Cancer Wisconsin dataset obtained from the UCI Machine Learning Repository. The research method adopts the CRISP-DM approach, which includes data collection, preprocessing, model development, testing, and evaluation stages. The preprocessing stage was carried out through data cleaning, data transformation, and data reduction to improve dataset quality before the modeling process. The testing process used split validation and evaluation based on accuracy, precision, recall, and Area Under Curve (AUC) metrics. The results indicate that the Neural Network algorithm achieved better performance than Decision Tree C4.5. Neural Network obtained an accuracy of 96.17%, precision of 95.80%, recall of 96.50%, and an AUC value of 0.989, which is categorized as excellent classification. Meanwhile, Decision Tree C4.5 achieved an accuracy of 93.50% and an AUC value of 0.945, categorized as very good classification. ROC Curve analysis demonstrates that Neural Network is more effective in distinguishing benign and malignant classes. Therefore, Neural Network is recommended as the best model to support early breast cancer detection based on machine learning, while Decision Tree C4.5 remains relevant for conditions requiring simpler and more interpretable models.]]>
Copyrights © 2024
