- Support Vector Machines (SVMs) create a boundary called a hyperplane to divide data points into partitions. The goal is to create a flat boundary that separates the data points as much as possible. - SVMs can be used for both classification and numeric prediction tasks. For classification, the SVM algorithm identifies a hyperplane that separates classes of data points. There may be multiple possible hyperplanes, and the algorithm aims to choose the best fitting one. - Non-linearly separable data can still be classified using SVMs by applying a cost value to data points that violate constraints, aiming to minimize total cost rather than find the maximum margin. Kernels can also be used to map the data into a higher