Sorry for the shilling, but here’s my upcoming project:
Please register your email address if you’re interested in it.
Edit: Sadly by the request of Coursera itself, I’ve removed this particular github repository. Coursera is doing the right thing though, don’t blame them.
I applied and followed Machine Learning course which is taught by Andrew Ng and offered freely online on Coursera (https://www.coursera.org/course/ml). However since the Octave installed on my computer doesn’t work properly (e.g. plotting functionality doesn’t work at all, etc), I decided to use python instead.
I managed to solve all of the programming exercises offered, in python instead of Octave. You can access the exercises’ solutions here: https://github.com/subokita/mlclass
In order to get started with python for machine learning, you might want to install these python libraries / modules:
- Numpy (http://www.numpy.org), for matrix and array manipulations
- Scipy (http://www.scipy.org), used mainly on optimization / minimization functions
- Matplotlib (http://matplotlib.org), for plotting
- NLTK (http://www.nltk.org), mainly for preprocessing text in spam detection section
- Sci-Kit Learn (http://scikit-learn.org/stable/), used for the SVM section. This is a great machine learning library
http://penandpants.com/2013/04/04/install-scientific-python-on-mac-os-x/ has a good write up on how to install most of these modules using Homebrew and pip.