This notebook was originally put together by [Jake Vanderplas](http://www.vanderplas.com) for PyCon 2014. [Peter Prettenhofer](https://github.com/pprett) adapted it for PyCon Ukraine 2014. Source and license info is on [GitHub](https://github.com/pprett/sklearn_pycon2014/).

An Introduction to scikit-learn: Machine Learning in Python

https://github.com/pprett/sklearn_pycon2014

Goals of this Tutorial

Introduce the basics of Machine Learning, and some skills useful in practice.
Introduce the syntax of scikit-learn, so that you can make use of the rich toolset available.

Schedule:

Outline:

10:00 - 10:30 Preliminaries: Setup & introduction

Making sure your computer is set-up
What is Machine Learning?
Quick review of Numpy and Matplotlib

10:30 - 11:15 Basic Principles of Machine Learning and the Scikit-learn Interface

Machine learning data layout
Supervised Learning
- Classification
- Regression
- Measuring performance
Unsupervised Learning
- Clustering
- Dimensionality Reduction
Evaluation of models
How to choose the right algorithm for your dataset

11:15 - 12:00 Supervised learning in-depth

Two important algorithms: Support Vector Machines and Random Forests
Application: recognizing handwritten digits

12:00 - 12:30 Unsupervised learning in-depth

Two important algorithms: PCA and K Means
Application: Eigen-Faces

12:30 - 13:00 Validation and Model Selection

Overfitting, Underfitting, bias, and variance
Improving your fit: validation curves and learning curves
Application: facial recognition

Preliminaries

This tutorial requires the following packages:

Python version 2.6-2.7 (Python 3.4 should work too)
numpy version 1.5 or later: http://www.numpy.org/
scipy version 0.9 or later: http://www.scipy.org/
matplotlib version 1.0 or later: http://matplotlib.org/
scikit-learn version 0.14 or later: http://scikit-learn.org
ipython version 1.0 or later, with notebook support: http://ipython.org

The easiest way to get these is to use an all-in-one installer such as Anaconda from Continuum. These are available for multiple architectures.

Anaconda Setup

If you're using Anaconda, simpy type

conda install scikit-learn

Otherwise it's best to install from source (requires a C compiler):

git clone https://github.com/scikit-learn/scikit-learn.git
cd scikit-learn
python setup.py install

Scikit-learn requires NumPy and SciPy, and examples require Matplotlib.

Note: some examples used in this tutorial require the scripts in the fig_code directory, which can be found within the notebooks subdirectory of the Github repository at https://github.com/pprett/sklearn_pycon2014/

Alternatives

Linux: If you're on Linux, you can use the linux distribution tools (by typing, for example apt-get install numpy or yum install numpy.
Mac: If you're on OSX, there are similar tools such as MacPorts or HomeBrew which contain pre-compiled versions of these packages.
Windows: Windows can be challenging: the best bet is probably to use a package installer such as Anaconda, above.

Checking your installation

You can run the following code to check the versions of the packages on your system:

(in IPython notebook, press shift and return together to execute the contents of a cell)



In [1]:

    
import numpy
print 'numpy:', numpy.__version__

import scipy
print 'scipy:', scipy.__version__

import matplotlib
print 'matplotlib:', matplotlib.__version__

import sklearn
print 'scikit-learn:', sklearn.__version__









    



numpy: 1.8.2
scipy: 0.9.0
matplotlib: 1.3.1
scikit-learn: 0.15.2

Useful Resources

scikit-learn: http://scikit-learn.org (see especially the narrative documentation)
matplotlib: http://matplotlib.org (see especially the gallery section)
IPython: http://ipython.org (also check out http://nbviewer.ipython.org)
Python Scientific Lecture Notes: http://scipy-lectures.github.io/ (best introduction to the scientific Python stack)
Andrew Ng's ML course: https://www.coursera.org/course/ml (also G. Hinton's Neural Network course)



In [ ]: