A whirlwind tour through matrix completion theory and methods, with exercises.
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
README.md
de-biasing_low-rank_projection.ipynb
matrix_completion_master.ipynb
mc_solve.py
mc_test.py
mc_util.py

README.md

Matrix Completion Whirlwind

By: Aaron Berk

About

This is a whirlwind introduction to matrix completion theory and methods. It was originally created for the 2017 BC Data Science Workshop. Some modifications have been made to protect proprietary data and heretofore unpublished methods.

This tutorial includes several exercises. Some of these exercises are novel — you coudl be the first to answer them!

Requirements

The notebook in this tutorial requires the cvxpy package, and relies on the splitting cone solver SCS, which is part of cvxgrp. For installation details, visit the cvxpy website, which should cover the automatic installation of SCS and other solvers.

This code also requires the latest versions of sklearn and scipy.

References

I have attempted to ensure that all appropriate references are given in-line in the notebook. If it seems I have made an omission, please e-mail me or submit a pull request.

Acknowledgements

I would like to thank Oscar Lopez for his help and comments with some of the finer points of this tutorial.