4.9 Support vector machines

A Support Vector Machine (SVM) is a decision-based prediction algorithm which can classify data into several groups. It is based on the concept of decision planes where thetraining data is mapped to a higher dimensional space and separated by a plane defining the two or more classes of data [1].

A simple example is seen in Figure 1 . Squares are data of class one while circles are data of class two. The SVMsets up the decision plane (in this case a simple line) and separates the two classes.

However, often the data is not distinguishable in two dimensions in which case it is mapped to higher dimensionsand the same process is done. An example is shown in Figure 2.

Support Vector Machine models can be classified into four major groups.

• C-SVM classification
• nu-SVM classification
• epsilon-SVM regression
• nu-SVM regression

The first two are classification algorithms which minimize different error functions while the second twoperform similar algorithms by regression [1].

[1] Chih-Chung Chang and Chih-Jen Lin, LIBSVM : a library for support vector machines, 2001. Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm