Written with StackEdit.
##Machine Learning with Scikit ###Setup and Installation
For this tutorial we will be working with a Python framework called Scikit Learn. This is a free machine learning library that will allow us to execute multiple ML techniques and methodologies.
Macs come pre-installed with Python, so let's dive right into it. (Windows users please pair programme)
Open up your terminal and enter:
$ which python
You should see:
/usr/local/bin/python
If you do not have Python, use Homebrew to install it:
$ brew install python
Also make sure that the version is 2.7. This will allow us to install the necessary dependencies.
$ python --version
Firstly, we need a package manager for python, to manage our libraries and dependencies. The most commonly used one is "pip". Most Macs come pre-installed with pip. Again you can check this using:
$ which pip
If you have it already, fantastic! Otherwise install it using EasyInstall.
$ easy_install pip
Now time for the dependencies. Before we can use Scikit Learn, we will need to have the following dependencies:
- Scipy - Scientific Computing Library
- Numpy - Library for complex mathematical computations
- matplotlib - Graph and Visuals plotting library
- Pandas - Data structures and Analysis
- iPython - Interactive console
- Sympy - Symbolic Mathematics
- virtualenv - Create virtual environments for python (Like Docker)
Start off by installing the virtual environment packages.
$ pip install virtualenv
$ pip install virtualenvwrapper
Now we can install Numpy.
$ pip install numpy
Before we can install Scipy, we need to get a compiler. Here we are using Fortran, an open source free compiler.
$ brew install gcc
$ pip install scipy
Now we can go ahead with MatplotLib, Pandas and iPython, along with its dependencies.
$ pip install matplotlib
$ pip install pandas
$ pip install jinja2
$ pip install tornado
$ pip install zmq
$ pip install ipython
$ pip install jupyter
We're almost there! One last package, our Scikit-Learn:
$ pip install -U scikit-learn
####Awesome! We are now ready for some Machine Learning with Scikit-Learn.