MITx-6.86x-version1

Shareable files for the MIT course Machine Learning with Python - From Linear Models to Deep Learning. The course took place from Feb to May 2020.

Project 4: Collaborative Filtering via Gaussian Mixtures

3. Expectation–maximization algorithm

The tests in test_3_em_algorithm.py use input and output from my submissions on the course for this section. These are used to check your results. Tests exist for the original toy dataset, and the estep, mstep, and run functions.

I've also added tests for the BIC function, which is in part 5.

I tested and developed these on Windows 10. Obviously, I can't guarantee that these will work on your computer.

If you don't want to run a particular test, you can either comment out the whole test or add the following above the function definition:

@unittest.skip('Skip this test')

7. Implementing EM for matrix completion

The tests in test_7_em_matrix_completion.py use input and output from my submissions on the course for this section for estep and mstep.

Also included are tests from the test_solutions.txt to create tests for estep, mstep and run.

Instructions

Files are shared between different parts. For convenience, I've put all the files in the same original git folder (part3) I created for the first tests.

1. Put all files in the same folder (netflix) as naive_em.py and em.py
2. To run the tests for each section, enter in a console:

python test_3_em_algorithm.py
python test_7_em_matrix_completion.py

The other files (*test_input*.py) contain the input and output test data.

Project 5: Text-based game

3. Q-learning algorithm

The tests in test.py use input from q_func.npy (a previously saved q_func array) to test functions tabular_q_learning() and epsilon_greedy() in agent_tabular_ql.py.

Save the files in the same folder as agent_tabular_ql.py and run:

python test.py