Improving Breast Cancer Classification Using Support Vector Classifier with Grid search CV for Optimal Hyperparameter Tuning

Breast cancer continues to be a major global health challenge, particularly for women, where early and reliable detection significantly improves patient outcomes. This research introduces a refined machine learning framework utilizing a Support Vector Classifier (SVC) in conjunction with GridSearchCV to optimize its hyperparameters for binary classification of breast tumors. The proposed approach systematically evaluates combinations of parameters such as kernel type, regularization constant (C), and gamma value, using 5-fold cross-validation. The optimized model was applied to the Wisconsin Diagnostic Breast Cancer dataset and achieved notable improvements across various evaluation metrics, including accuracy, recall, precision, and F1-score. Results demonstrate the efficacy of hyperparameter tuning in enhancing predictive performance, validating the method’s potential as a reliable diagnostic support tool for early cancer screening.