- DenseSVMTrainer takes training data and builds a predictive model in binary format.
- DenseSVMPredictor uses the model to predict the class of each sample in a test data set.
- DenseSVMModelPrinter displays readable information about the model.
The DenseSVMTrainer and DenseSVMPredictor functions are designed for input in dense format; that is, each table column contains values of a single attribute and there is a single row for each sample (observation).
This implementation of DenseSVM function includes a linear SVM based on a Pegasos algorithm and a non-linear SVM based on the Hash-SVM model described in the paper "Hash-SVM: Scalable Kernel Machines for Large-Scale Visual Classification," by Yadong Mu, Gang Hua, Wei Fan, and Shih-Fu Chang (http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6909525).