This repo reimplements the programming assignments from Andrew Ng's Cousera Machine Learning course. The assignment pdf files and data files are Coursera's and under their control. All the python code is mine and may be reproduced, altered and used for any purpose, with attribution. Although you can probably find more efficient implementations elsewhere.
In each exercise directory there is an IPython Notebook file (.ipynb) which contains the code and in-line documentation. Github will render the IPython Notebooks for viewing in a browser (although Github seems to have trouble with rendering LaTax). Alternative ways to view the code include using the official Jupyter Notebook viewwer or downloading files from this repo. If downloading, I recommend creating a virtual environment and pip installing the dependencies listed in the requirements.txt file into the VirtualEnv. Each directory also includes that week's PDF assignment file for reference, and any needed data files.
-
ex1: Linear regression, gradient descent, parabolic error surface, iso-error curves, feature scaling & normalization
-
ex2: Logistic regression, optimization with SciPy, regularization, creating polynomial features, plotting non-linear decision boundaries.