liquidSVM: A Fast and Versatile SVM Package

Support vector machines (SVMs) and related kernel-based learning algorithms are a well-known class of machine learning algorithms, for non-parametric classification and regression. liquidSVM is an implementation of SVMs whose key features are: fully integrated hyper-parameter selection, extreme speed on both small and large data sets, full flexibility for experts, and inclusion of a variety of different learning scenarios: multi-class classification, ROC, and Neyman-Pearson learning, and least-squares, quantile, and expectile regression.

Version: 1.2.1
Depends: R (≥ 2.12.0), methods
Suggests: knitr, rmarkdown, deldir, testthat
Enhances: mlr, ParamHelpers
Published: 2017-07-19
Author: Ingo Steinwart, Philipp Thomann
Maintainer: Philipp Thomann <philipp.thomann at>
License: AGPL-3
Copyright: Ingo Steinwart, Philipp Thomann, Mohammad Farooq
NeedsCompilation: yes
Citation: liquidSVM citation info
Materials: README
CRAN checks: liquidSVM results


Reference manual: liquidSVM.pdf
Vignettes: liquidSVM Demo
liquidSVM Documentation
Package source: liquidSVM_1.2.1.tar.gz
Windows binaries: r-devel:, r-release:, r-oldrel:
OS X binaries: r-release: liquidSVM_1.2.1.tgz, r-oldrel: liquidSVM_1.2.1.tgz
Old sources: liquidSVM archive


Please use the canonical form to link to this page.