Support Vector Machines - SVM Tab

Select the SVM tab of the Support Vector Machines dialog box to access the options described here.

SVM type

In this group box, select the type of SVM model you want to use for the analysis. This group box further divides into three sub-groups:

Classification SVM

For classification problems, select one of the option buttons in this group box to specify the SVM model of your choice. You can select either Type 1 or Type 2. According to which type is selected, options will become available in the Training constants group box (see below).

Regression SVM

For regression problems, select one of the option buttons in this group box to specify the SVM model of your choice. You can select either Type 1 or Type 2. According to which type is selected, options will become available in the Training constants group box (see below).

Training constants

Depending on the type of SVM model you have chosen (see above), use the fields in this group box (when applicable) to specify the value for the training parameters. Although these parameters are unknown a priori, you can estimate and specify their values. Alternatively, you can use the cross-validation algorithm (see the documentation for the Cross-validation tab) to obtain estimates of their values. The Training constants parameters include:

Capacity. Applicable to Classification SVM Type 1, Regression SVM Type 1, and Regression SVM Type 2.

Epsilon. Applicable to Regression SVM Type 1.

Nu. Applicable to Classification SVM Type 2 and Regression SVM Type 2.