The LibCVM Toolkit is a C++ implementation of the improved Core Vector Machine (CVM) and recently developed Ball Vector Machine (BVM), which are fast Support Vector Machine (SVM) training algorithms using core-set approximation on very large scale data sets. It is adapted from the LIBSVM implementation (version 2.85). The code has been used on a large range of problems, including network intrusion detection, face detection and implicit surface modeling. The main features of the toolkit are the following: more stable core set selection adaptive epsilon approximation solution sparse caching of kernel evaluations can handle millions of training examples supports standard and several other nonlinear kernel functions supports dense and sparse vector representations supports multiple training data files and label renaming supports BVM/CVM/CVM-LS for large-scale classification supports Core Vector Data Description (CVDD) for large-scale novelty detection supports Core Vector Regression (CVR) for large-scale sparse least-squares regression Pending features: multiclass CVM/BVM probabilistic output estimation for CVM If you have any suggestions or bug findings, please email to Thank you!

Keywords for this software

Anything in here will be replaced on browsers that support the canvas element