liquidSVM v1.2.2

0

Monthly downloads

0th

Percentile

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.

Readme

liquidSVM

liquidSVM is a package written in C++ that provides SVM-type solvers for various classification and regression tasks. Because of a fully integrated hyper-parameter selection, very carefully implemented solvers, multi-threading and GPU support, and several built-in data decomposition strategies it provides unprecedented speed for small training sizes as well as for data sets of tens of millions of samples.

You can use it e.g. for multi-class classification, least squares (kernel) regression, or even quantile regression, etc.:

install.packages("liquidSVM")
library(liquidSVM)

model <- mcSVM(Species ~ ., iris)
predict(model, iris)

model <- lsSVM(Height ~ ., trees)
y <- predict(model, trees)

model <- svmQuantileRegression(Height ~ ., trees)
y <- test(model, trees)

If you install build the package to be used on several machines please use the following:

install.packages("liquidSVM", configure.args="generic")

For details please look at the vignettes demo and documentation. Also check the help ?liquidSVM and ?svm. For the command-line version and other bindings go to (http://www.isa.uni-stuttgart.de/software/).

Functions in liquidSVM

Name Description
bsSVM Bootstrap
Configuration liquidSVM model configuration parameters.
rocSVM Receiver Operating Characteristic curve (ROC curve)
reg-1d reg-1d.train and reg-1d.test
getCover Get Cover of partitioned SVM
getSolution Retrieve the solution of an SVM
banana banana-bc.train, banana-bc.test banana-mc.train, and banana-mc.test
nplSVM Neyman-Pearson-Learning
init.liquidSVM Initialize an SVM object.
plotROC Plots the ROC curve for a result or model
kern Calculates the kernel matrix.
svm Convenience function to initialize, train, select, and optionally test an SVM.
test.liquidSVM Tests new data using the selected SVM.
mlr-liquidSVM liquidSVM functions for mlr
print.liquidSVM Printing an SVM model.
clean.liquidSVM Force to release the internal memory of the C++ objects associated to this model.
mcSVM Multiclass Classification
predict.liquidSVM Predicts labels of new data using the selected SVM.
trainSVMs Trains an SVM object.
selectSVMs Selects the best hyper-parameters of all the trained SVMs.
write.liquidData Write Smldata
setDisplay Set display info mode that controls how much information is displayed by liquidSVM C++ routines. Usually you will use display=d in svm(...) etc.
read.liquidSVM Read and Write Solution from and to File
qtSVM Quantile Regression
errors Obtain the test errors result.
command-args liquidSVM command line options
exSVM Expectile Regression
compilationInfo Compilation information: whether the library was compiled using SSE2 or even AVX.
liquidSVM-package liquidSVM for R
lsSVM Least Squares Regression
liquidData Loads or downloads training and testing data
liquidSVM-class A Reference Class to represent a liquidSVM model.
No Results!

Vignettes of liquidSVM

Name
demo_cache/banana-bc-roc-ls.fsol
demo_cache/banana-mc.fsol
demo_cache/reg.fsol
demo_cache/result_et.R
demo_cache/result_ex.R
demo_cache/result_npl.R
demo_cache/result_qt.R
demo_cache/result_roc.R
md/global-and-grid.md
demo.Rmd
documentation.Rmd
No Results!

Last month downloads

Details

Type Package
Date 2019-01-10
Copyright Ingo Steinwart, Philipp Thomann, Mohammad Farooq
URL https://github.com/liquidSVM/liquidSVM
License AGPL-3
VignetteBuilder knitr
RoxygenNote 6.0.1
NeedsCompilation yes
Packaged 2019-01-10 19:28:51 UTC; philippthomann
Repository CRAN
Date/Publication 2019-01-10 20:10:03 UTC

Include our badge in your README

[![Rdoc](http://www.rdocumentation.org/badges/version/liquidSVM)](http://www.rdocumentation.org/packages/liquidSVM)