Tools for use of the SVM package.
This provides practical programs for using SVMs to classify or regress real data. It also contains graphical demonstrations to explain how SVM works.
Interface Summary Interface Description SuffixTreeKernel.DepthScalerEncapsulates the scale factor to apply at a given depth.
Class Summary Class Description ClassifierExampleA simple toy example that allows you to put points on a canvas, and find a polynomial hyperplane to seperate them. ClassifierExample.PointClassifierAn extention of JComponent that contains the points & encapsulates the classifier. Classify SuffixTreeKernelComputes the dot-product of two suffix-trees as the sum of the products of the counts of all nodes they have in common. SuffixTreeKernel.MultipleScalarScale using a multiple of two DepthScalers. SuffixTreeKernel.NullModelScalerScales by 4^depth - equivalent to dividing by a probablistic flatt prior null model SuffixTreeKernel.SelectionScalarScale using a BitSet to allow/disallow depths. SuffixTreeKernel.UniformScalerScale all depths by 1.0 SVM_Light SVM_Light.LabelledVector Train TrainRegression