Support Vector Machine

This chapter provides a comprehensive overview of support vector machines (SVM), a critical algorithm in classification and regression analysis. It begins with a basic introduction to SVM, including its concept, application in binary classification, and the significance of support vectors. The chapter then delves into the comparison of SVM with logistic regression, focusing on their differing sensitivities to data through their respective loss functions. A substantial portion of the chapter is dedicated to explaining the role of the kernel in SVM, particularly the radial basis function (RBF) kernel, and its capability to transform data for effective classification. The chapter also discusses the tuning of hyperparameters in SVM, emphasizing the importance of the C parameter and the influence of the RBF kernel’s gamma parameter. Practical applications of SVM are explored through a case study on breast cancer diagnosis and the Titanic Kaggle dataset, highlighting data preprocessing techniques like one-hot encoding, normalization, and model scoring methods.