mix-SQP experiments

Demos for mix-SQP paper

These webpages contain demos of mix-SQP, a fast algorithm for maximum-likelihood estimation of mixture proportions, and examples comparing mix-SQP against other alternative approaches, such as EM and interior-point methods.

The webpages were generated by rendering the Jupyter notebooks into webpages using nbconvert.

For more information, including Setup Instructions, please see the README in the git repository.

These are the demonstrations we have implemented:

  1. Illustration of mix-SQP solver applied to a small data set, and a large one.

  2. mix-SQP demo with different low-rank matrix approximations.

  3. Comparing performance and accuracy of EM, IP and mix-SQP algorithms.

  4. Profiling adaptive shrinkage computations with mix-SQP and IP solvers.

  5. Comparing performance of active-set and interior-point methods for solving quadratic subproblem inside SQP.

Credits

This project was developed by Youngseok Kim, Peter Carbonetto, Matthew Stephens and Mihai Anitescu at the University of Chicago.

© 2017-2018 Youngseok Kim, Peter Carbonetto, Matthew Stephens & Mihai Anitescu.