Robust Linear Regression

Code Organization

The idea behind the code is based on having functions that let you preform the simulations that you want to use. The principal files are the following:

• fpeqs.py contains all of the functions to run simulations for the fixed point equations
• numerics.py contains the functions to run the numerical experiments
• numerical_functions.py mostly contains all the functions to perform numerical integrations and that are called from fpeqs.py.
• utils.py contains functions to load and save simulations data
• integration_utils.py contains functions that are used to perform integrations

All of the functions used to save and load files are contained in src.py.

The folder src_cluster contains all files that only reference themselves and present a parallel implementation. These files are also working on minimal dependances (numpy, scipy, sklearn) and are meant to be runned in parallel.

Code usability

An example of code is presented in the file together_theory_experiment.py. One can create a dictionary with all the parameters of the simulations and then call the routine in utils that runs and saves the experiment.

Tests

To run the tests for the code one should run the file unittests/run_tests.py.