support vector machine SVM

Support Vector Machines (SVMs):

 

 These are the most recent supervised machine learning technique.

 Support Vector Machine (SVM) models are closely related to classical multi layer perceptron neural networks.

SVMs revolve around the notion of a margin either side of a hyperplane that separates two data classes.

 Maximizing the margin and thereby creating the largest possible distance between the separating hyperplane and the instances on either side of it has been proven to reduce an upper bound on the expected generalization error.