<![CDATA[Breast cancer is one of the leading causes of death among women worldwide. Early detection of breast cancer, whether benign or malignant, is crucial for increasing the chances of successful treatment. This study aims to develop a breast cancer prediction model using the Support Vector Machine (SVM) method combined with Recursive Feature Elimination (RFE) to enhance prediction efficiency and accuracy. The RFE method is applied to select significant features that contribute substantially to breast cancer classification. The dataset used in this study is derived from the Breast Cancer Wisconsin Diagnostic and tested under various feature selection scenarios. The results indicate that features such as area_worst, texture_worst, radius_worst, concave_points_mean, and concavity_mean consistently emerge as the most relevant. The SVM model with RFE, utilizing 10 features, achieves the best accuracy of 98.25%, with an optimal balance among precision, recall, and F1-score. This approach effectively filters relevant information, reduces noise from insignificant features, and improves data efficiency and interpretability. These findings affirm that the combination of SVM and RFE can serve as an effective method for breast cancer prediction and provide a robust foundation for future clinical analyses.]]>
Copyrights © 2024
