Top Page | Upper Page | Contents | About This Site | JAPANESE

Support Vector Machine (SVM)

The use of Support Vector Machine (SVM) is similar to Discriminant Analysis and Logistic Regression Analysis. It also uses a line (super plane) to classify samples.

Maximization of the Margin

Support Vector Machine uses "margin." It is a function to evaluate the generality of the model.

It is used to avoid over-fitting and local optimization.

Support Vector

The formulation of SVM uses the all data. But the formulation of output includes only the data near the line. The data called "Support Vector".

The output of RapidMiner tells the information of support vectors.

Kernel Method

There are cases that a simple line is not valid to classify. We can use simple line method by the help of Kernel method.

Kernel method is made in the studies of SVM. But this method also used for other multi-variable analysis methods.

Difference of Good Distribution by Methods

Difference of Output by Methods

Selection of Methods

NEXT Kernel Method