Python

# Learning From Data

The code for the homework assignments for the Cal Tech CS 1156x class offered on edX.org. The class was about Machine Learning, focusing on the theory. My class notes are also available. I decided to do the assignments in Python to gain some experience with NumPy, which is quickly becoming the industry choice for scientific computing. I had experience with R, but I preferred to use a "real" language.

I created a simple testing mechanism to automate testing of the experiments that were developed, allowing me to do something like:

# constructor is Question(question_str, choices, answer)
question8 = Question("8. in sample error",
                     [0, 0.1, 0.3, 0.5, 0.8], 'd')

in_sample_error = experiment()
question8.check(in_sample_error)

Which would output something like:

The homework assignments covered many topics such as the simple perceptron learning algorithm, support vector machines, logistic regression via stochastic gradient descent, regularization, and validation.

August 27, 2015