An In-Depth Analysis of Robustness to Noise in Soft Margin Support Vector Machines

This study investigates the resilience of the Soft Margin Support Vector Machine (SVM) in the presence of noise. Soft Margin SVMs are well-known for their adaptability when dealing with datasets that challenge perfect linear separability. This in-depth analysis investigates the robustness of Soft Margin Support Vector Machines (SVMs) in the presence of noise within training data. Recognizing the paramount importance of SVMs in machine learning applications, especially in real-world scenarios where data may be inherently noisy, this study contributes to understanding the performance of soft margin formulations. Employing a carefully defined noise model, the evaluation explores various noise levels and types, employing standard metrics such as accuracy, precision, recall and F1-Score. The experimental setup involves diverse datasets and systematically tuned SVM parameters. Results reveal the response of soft margin SVMs to different noise conditions,