This documentation is for scikit-learn version 0.11-gitOther versions


If you use the software, please consider citing scikit-learn.

This page

1. Installing scikit-learn

There are different ways to get scikit-learn installed:

  • Install the version of scikit-learn provided by your operating system distribution . This is the quickest option for those who have operating systems that distribute scikit-learn.
  • Install an official release. This is the best approach for users who want a stable version number and aren’t concerned about running a slightly older version of scikit-learn.
  • Install the latest development version. This is best for users who want the latest-and-greatest features and aren’t afraid of running brand-new code.


If you wish to contribute to the project, it’s recommended you install the latest development version.

1.1. Installing an official release

1.1.1. Installing from source

Installing from source requires you to have installed python (>= 2.6), numpy (>= 1.3), scipy (>= 0.7), setuptools, python development headers and a working C++ compiler. Under Debian-based systems you can get all this by executing with root privileges:

sudo apt-get install python-dev python-numpy python-numpy-dev python-setuptools python-numpy-dev python-scipy libatlas-dev g++


In Order to build the documentation and run the example code contains in this documentation you will need matplotlib:

sudo apt-get install python-matplotlib


On Ubuntu LTS (10.04) the package libatlas-dev is called libatlas-headers Easy install

This is usually the fastest way to install the latest stable release. If you have pip or easy_install, you can install or update with the command:

pip install -U scikit-learn


easy_install -U scikit-learn

for easy_install. Note that you might need root privileges to run these commands. From source package

Download the package from , unpack the sources and cd into archive.

This packages uses distutils, which is the default way of installing python modules. The install command is:

python install

1.1.2. Windows installer

You can download a windows installer from downloads in the project’s web page. Note that must also have installed the packages numpy and setuptools.

This package is also expected to work with python(x,y) as of

Installing on Windows 64bit

To install a 64bit version of the scikit, you can download the binaries from Note that this will require a compatible version of numpy, scipy and matplotlib. The easiest option is to also download them from the same URL.

1.1.3. Building on windows

To build scikit-learn on windows you will need a C/C++ compiler in addition to numpy, scipy and setuptools. At least MinGW (a port of GCC to Windows OS) and the Microsoft Visual C++ 2008 should work out of the box. To force the use of a particular compiler, write a file named setup.cfg in the source directory with the content:



where my_compiler should be one of mingw32 or msvc.

When the appropriate compiler has been set, and assuming Python is in your PATH (see Python FAQ for windows for more details), installation is done by executing the command:

python install

To build a precompiled package like the ones distributed at the downloads section, the command to execute is:

python bdist_wininst -b doc/logos/scikit-learn-logo.bmp

This will create an installable binary under directory dist/.

1.2. Third party distributions of scikit-learn

Some third-party distributions are now providing versions of scikit-learn integrated with their package-management systems.

These can make installation and upgrading much easier for users since the integration includes the ability to automatically install dependencies (numpy, scipy) that scikit-learn requires.

The following is a list of Linux distributions that provide their own version of scikit-learn:

1.2.1. Debian and derivatives (Ubuntu)

The Debian package is named python-sklearn (formerly python-scikits-learn) and can be installed using the following commands with root privileges:

apt-get install python-sklearn

Additionally, backport builds of the most recent release of scikit-learn for existing releases of Debian and Ubuntu are available from NeuroDebian repository .

1.2.2. Python(x, y)

The Python(x, y) distributes scikit-learn as an additional plugin, which can be found in the Additional plugins page.

1.2.3. Enthought Python distribution

The Enthought Python Distribution already ships a recent version.

1.2.4. Macports

The macport’s package is named py26-sklearn and can be installed by typing the following command:

sudo port install py26-scikits-learn

1.2.5. NetBSD

scikit-learn is available via pkgsrc-wip:

1.3. Bleeding Edge

See section Retrieving the latest code on how to get the development version.

1.4. Testing

Testing requires having the nose library. After installation, the package can be tested by executing from outside the source directory:

python -c "import sklearn; sklearn.test()"

This should give you a lot of output (and some warnings) but eventually should finish with the a text similar to:

Ran 601 tests in 27.920s

otherwise please consider posting an issue into the bug tracker or to the Mailing List.

scikit-learn can also be tested without having the package installed. For this you must compile the sources inplace from the source directory:

python build_ext --inplace

Test can now be run using nosetest:

nosetests sklearn/

If you are running the development version, this is automated in the commands make in and make test.


Because nosetest does not play well with multiprocessing on windows, this last approach is not recommended on such system.