Support Vector Machines

From Machine Learning
Jump to: navigation, search

Support Vector Machines (SVM) work by finding an optimal hyperplane with maximum boundaries between datasets[1]. The image below illustrates this concept – on the left the green lines are all the possible separators, and the SVM finds the maximum margin between both datasets as shown on the right. The term “support vector” refers to data points closer to the hyperplane that influence the position and orientation of the hyperplane [2].

SVM.png

An SVM can be used in classification or regression problems, though it is more commonly used for classification problems.

Given a set of training examples, each marked as belonging to one or the other of two categories, an SVM training algorithm builds a model that assigns new examples to one category or the other, making it a non-probabilistic binary linear classifier.

The hyperplane can be a line (2D, when there are only two input features) or a plane (3D, when there are 3 input features).

Hyperplanes.png

The important hyperparameter to consider in SVM is C: the margin classifier that determines the margin between the hyperplane and the datapoints. A larger value of C can make SVM more sensitive to outliers and more susceptible to overfitting. A smaller value of C can provide a more robust model

References