Coursera Machine Learning class programming exercises
Python versions of Professor Ng's machine learning algorithms could make them easier to integrate into web platforms and integration with large databases.
The original algorithms were presented in videos with mathematical notation. The programming exercises and tutorials focused on using Octave (or Matlab) to implement solutions.
These python versions rely on numpy, so they may not be too handy for deploying on heroku web servers ;)
- feature normalization
- linear regression
- polynomial regression
- regression (fitting) to any arbitrary linear combination of features or functions
- logistic regression (classification)
Various quizzes and programming exercises as example usages.