Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/liquidSVM/liquidSVM

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.
https://github.com/liquidSVM/liquidSVM

apache-spark c-plus-plus classification expectile-regression machine-learning matlab ml octave python quantile-regression r r-package regression rstats svm

Last synced: 11 days ago
JSON representation

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.

Awesome Lists containing this project

README 

        

## General Information



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,

* Bindings for [R](#R), [Python](#python), [MATLAB / Octave](#matlab-octave), [Java](#java), and [Spark](#spark),

* full flexibility for experts, and

* inclusion of a variety of different learning scenarios:

    - multi-class classification, ROC, and Neyman-Pearson learning,

    - least-squares, quantile, and expectile regression.



For questions and comments just contact us via

[mail](http://www.uni-stuttgart.de/cgi-bin/mail.cgi?liquidSVM=mathematik.uni-stuttgart.de).

There you also can ask to be registerd to our mailing list.



liquidSVM is licensed under [AGPL 3.0](http://www.gnu.org/licenses/agpl-3.0.html). In case you need another license, please contact [me](http://www.isa.uni-stuttgart.de/Steinwart/).



## Command Line interface



[Installation instructions](http://www.isa.uni-stuttgart.de/software/install.txt) for the command line versions.



|                                      |                                                                                    |

|--------------------------------------|------------------------------------------------------------------------------------|

| Terminal version for Linux/OS X      | [liquidSVM.tar.gz](http://www.isa.uni-stuttgart.de/software/liquidSVM.tar.gz)      |

| Terminal version for Windows (64bit) | avx2: [liquidSVM.zip](http://www.isa.uni-stuttgart.de/software/avx2/liquidSVM.zip) |

|                                      | avx:  [liquidSVM.zip](http://www.isa.uni-stuttgart.de/software/avx/liquidSVM.zip)  |

|                                      | sse2: [liquidSVM.zip](http://www.isa.uni-stuttgart.de/software/sse2/liquidSVM.zip) |

| Previous versions                    | [v1.1](v1.1) (June 2016), [v1.0](v1.0) (January 2016)                              |



On Linux and Mac on the terminal `liquidSVM` can be used in the following way:

```bash

wget www.isa.uni-stuttgart.de/software/liquidSVM.tar.gz

tar xzf liquidSVM.tar.gz

cd liquidSVM

make all

scripts/mc-svm.sh banana-mc 1 2

```



## R



Read the [demo vignette](http://www.isa.uni-stuttgart.de/software/R/demo.html) for a tutorial on installing liquidSVM-package and how to use it and the [documentation vignette](http://www.isa.uni-stuttgart.de/software/R/documentation.html) for more advanced installation options and usage.



An easy usage is:

```r

install.packages("liquidSVM")

library(liquidSVM)

banana <- liquidData('banana-mc')

model <- mcSVM( Y~. , banana$train, display=1, threads=2)

result <- test(model, banana$test)

errors(result)

```



## Python



Read the [demo notebook](http://www.isa.uni-stuttgart.de/software/python/demo.html) for a tutorial on installing liquidSVM-package and how to use it and the [homepage](bindings/python/) for more advanced installation options and usage.



To install use:

```bash

pip install --user liquidSVM

```

and then in Python you can use it e.g. like:

```python

from liquidSVM import *

banana = LiquidData('banana-mc')

model = mcSVM(banana.train, display=1, threads=2)

result, err = model.test(banana.test)

```



## MATLAB / Octave



The [MATLAB bindings](bindings/matlab/) are currently getting a better interface,

and this is a preview version.



> It does currently not work on Windows.



For installation download the Toolbox

[liquidSVM.mltbx](http://www.isa.uni-stuttgart.de/software/matlab/liquidSVM.mltbx)

and install it in MATLAB by double clicking it.

To compile and add paths issue:

```matlab

makeliquidSVM native

```

Then you can use it like:

```matlab

banana = liquidData('banana-mc');

model = svm_mc(banana.train, 'DISPLAY', 1, 'THREADS', 2);

[result, err] = model.test(banana.test);

```



Most of the code also works in `Octave`

if you use [liquidSVM-octave.zip](http://www.isa.uni-stuttgart.de/software/matlab/liquidSVM-octave.zip).



## Java

The main homepage is [here](bindings/java/).

For installation download [liquidSVM-java.zip](http://www.isa.uni-stuttgart.de/software/java/liquidSVM-java.zip) and unzip it.

The classes are all in package `de.uni_stuttgart.isa.liquidsvm` and an easy example is:

```java

LiquidData banana = new LiquidData("banana-mc");

SVM model = new MC(banana.train, new Config().display(1).threads(2));

ResultAndErrors result = model.test(banana.test);

```

If this is implemented in the file `Example.java` this can be compiled and run using

```bash

# if you want to compile the JNI-native library:

make lib

# compile your Java-Code

javac -classpath liquidSVM.jar Example.java

# and run it

java -Djava.library.path=. -cp .:liquidSVM.jar Example

```



## Spark

This is a preview version, see [Spark](bindings/spark/) for more details.

Download [liquidSVM-spark.zip](http://www.isa.uni-stuttgart.de/software/spark/liquidSVM-spark.zip) and unzip it.

Assume you have `Spark` installed in `$SPARK_HOME` you can issue:

```bash

make lib

export LD_LIBRARY_PATH=.:$LD_LIBRARY_PATH

$SPARK_HOME/bin/spark-submit \

  --class de.uni_stuttgart.isa.liquidsvm.spark.App \

  liquidSVM-spark.jar banana-mc

```

If you have configured `Spark` to be used on a cluster with `Hadoop` use:

```bash

hdfs dfs -put data/covtype-full.train.csv data/covtype-full.test.csv .

make lib

$SPARK_HOME/bin/spark-submit --files ../libliquidsvm.so \

  --conf spark.executor.extraLibraryPath=. \

  --conf spark.driver.extraLibraryPath=. \

  --class de.uni_stuttgart.isa.liquidsvm.spark.App \

  --num-executors 14 liquidSVM-spark.jar covtype-full

```



Extra Datasets for the Demo

---------------------------



[covertype data set with 35.090 training and 34.910 test samples](http://www.isa.uni-stuttgart.de/software/covtype.35000.zip)



[covertype data set with 522.909 training and 58.103 test samples](http://www.isa.uni-stuttgart.de/software/covtype.zip)



Both datasets were compiled from [LIBSVM's version of the covertype dataset](https://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/binary.html), which

in turn was taken from the [UCI repository](http://mlr.cs.umass.edu/ml/datasets/Covertype) and preprocessed as in [\[RC02a\].](https://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/ref.html#RC02a)

Copyright for this dataset is by Jock A. Blackard and Colorado State University.



Citation

--------



If you use liquidSVM, please cite it as:



> I. Steinwart and P. Thomann.

> *liquidSVM: A fast and versatile SVM package.*

> [*ArXiv e-prints 1702.06899*](http://arxiv.org/abs/1702.06899), February 2017.