This package provides the most basic concord-ista algorithm for computing the concord estimate in python3. The algorithm is based on the paper by Oh, Dalal, Khare and Rajaratnam.
Prerequisite for installing this package is the Eigen linear algebra library in addition to numpy and scipy. In order to install this package, issue the following command:
pip install pyconcordTo test if everything is fine, start python interpreter and execute something like the following:
from concord import concord
import numpy as np
x = np.random.randn(13, 9)
omega = concord(x, 0.3)
print(np.round(omega.todense(),2))Should work on linux and mac. Not compiled on Windows platforms.
If you want to run the tests or develop, run:
make develop
pytest -s