Exploring SVM hyperparameters
Support Vector Machine (SVM) is an ML model that utilizes lines or hyperplanes, along with some linear algebra transformations, to perform a classification or regression task. All the algorithms discussed in the previous sections can be classified as tree-based algorithms, while SVM is not part of the tree-based group of ML algorithms. It is part of the distance-based group of algorithms. We usually called the linear algebra transformation in SVM a kernel. This is responsible for transforming any problem into a linear problem.
The most popular and well-maintained implementation of SVM in Python can be found in the scikit-learn package. It includes implementations for both regression (SVR
) and classification (SVC
) tasks. Both of them have very similar hyperparameters with only a few small differences. The following are the most important hyperparameters for SVM, starting with the most important to the least based on their impact on model performance...