A Comparative Performance Analysis of SVM Kernels in Automated Breast Cancer Diagnosis

This research evaluates the effectiveness of Support Vector Machines (SVMs) in automated breast cancer diagnosis, focusing on the comparative performance of three kernel functions: Linear, Polynomial, and Radial Basis Function (RBF). Using the Breast Cancer Wisconsin (Diagnostic) Dataset, we preprocessed data and applied SVM models with hyperparameter tuning. The Linear kernel yielded the highest accuracy of 98.83%, with balanced precision and recall, making it suitable for linearly separable data. The RBF kernel also achieved high performance, with an accuracy of 97.66% and an F1-score of 96.83%, demonstrating proficiency in handling non-linear relationships. In contrast, the Polynomial kernel initially showed lower recall but improved significantly to 98.25% accuracy after tuning. These findings underscore SVM’s potential in reliable, automated tumor classification, contributing to more accurate breast cancer diagnosis.