SVM Demo

I provided an implementation of classification (binary and one-class) and regression SVM for the Text Garden library. It is written entirely in C++ and is based on decomposition method as proposed by J. Thorsten in Making Large-Scale SVM Practical. It was implemented with efficiency and expandability in mind and new algoriths are easy to add.

I also made a little graphical toy example. It shows the working of the binary SVM classifier in two dimension using linear, polynomial or Gaussian kernel. The SVM engine used is the one from Text Garden. You can download the SVM demo. The program does not require any utilities from the Text Garden. However, .NET Framework 1.1 is needed.