This is the project page for a project that I am proposing for Google's Summer of Code 2006. After a very successful first venture into machine learning for a Robotics class (project page here),  I was intrigued to learn more and implement other algorithms in Python. Unfortunately, I found that much of what is currently available online falls into one of three problematic categories:

• Not well-documented
• Very application-specific
• Scattered

This makes it difficult for many people to use Python to implement machine learning, or artificial intelligence, into their software. Indeed, I found a number of repeated projects across the web; people repeating others' work because they didn't find it or they didn't know how to use it.

What I propose is to first bring together all the work that has been done, document it extensively so that it is useful to everyone, and create a standard Python library that is easy to use, robust, and portable. After this is done, it will be clear what methods of machine learning have not been implemented in Python code yet, or what needs to be modified for better ease of use. I would then fill in these holes with additional well-documented libraries of Python code.

This is a very cursory description of the project, and more details can be found through the links at left. Please don't hesitate to contact me with any questions.