Skip to main content

A package to solve low rank matrix completion problems

Project description

CompleteThat (v0.1dev)

CompleteThat is a python package that solves the low rank matrix completion
problem. Given a low rank matrix with partial entries the package solves an
optimization problem to estimate the missing entries.

Mathematically, the package solves a relaxation (using the nuclear norm or the
Frobenius norm of the objective matrix) of the following problem:

minimize_{X} ||X||
st. X(i,j) = M(i,j) \forall (i,j)\in \Omega,
where, M represents the data matrix and \Omega represents the set of p
observed entries of M


>>> from completethat import MatrixCompletion
>>> problem = MatrixCompletion(M)
>>> problem.complete_it(algo_name)
>>> X = problem.get_matrix()
>>> out_info = problem.get_out()

>>> from completethat import MatrixCompletionBD
>>> problem = MatrixCompletionBD('input_data.txt')
>>> problem.train_sgd(dimension=6,init_step_size=.01,min_step=.000001, reltol=.001,rand_init_scale=10, maxiter=1000,batch_size_sgd=50000,shuffle=True)
>>> problem.validate_sgd('test_data.txt')
>>> problem.save_model()


This package was written by Joshua Edgerton and Esteban Fajardo


This package is the result of the final project for the class EEOR E4650: Convex
Optimization at Columbia University, Fall 2014. We would like to thank the
authors of the different algorithms used in the package to solve the problem.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

completethat-0.1dev.tar.gz (5.9 kB view hashes)

Uploaded source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page