Optimization of SVM Parameters for Accurate Breast Cancer Diagnosis

It is termed Breast Cancer, and it is the most popular type of cancer among women throughout the world. Early detection is essential to treat and save a patient's life properly. Although kernel - based support vector machine (SVM) has achieved high efficacy in diagnosing breast cancer, its performance depends on how to set its parameters. Therefore, it is important to optimize the SVM parameters for accurate diagnosis. This research presents a methodical framework for rational SVM selection in breast cancer diagnosis. The approach to this novel method uses the grid search technique in tuning two important parameters: the regularization part, which is the term, and the kernel coefficient. We also investigate the impact of kernel function choice on SVM performance. The experiments show that the proposed optimized SVM parameters can achieve higher accuracy in breast cancer diagnosis than conventional (default) SVM parameter settings. On the other hand, kernel function, which is another significant characteristic of classification performance, is also affected by this choice. This method can help medical professionals identify breast cancer with greater precision, resulting in timely and effective treatments. Further, this can also be used for other classification tasks where SVM is used.