QUIC and BigQUIC
- The QUadratic Inverse Covariance algorithm
The QUadratic Inverse Covariance algorithm (latest release 1.1) implements the l1 regularized Gaussian maximum likelihood estimation of the inverse of a covariance matrix.
Download
Citation
This software is released under the GNU License but please acknowledge its use with a citation to at least one of the following publications:- QUIC: Quadratic Approximation for Sparse Inverse Covariance Matrix Estimation (pdf, software)
C. Hsieh, M. Sustik, I. Dhillon, P. Ravikumar.
Journal of Machine Learning Research (JMLR), October 2014. - BIG & QUIC: Sparse Inverse Covariance Estimation for a Million Variables (pdf, software)
C. Hsieh, M. Sustik, I. Dhillon, P. Ravikumar, R. Poldrack.
In Neural Information Processing Systems (NIPS), December 2013. (Oral) - A Divide-and-Conquer Method for Sparse Inverse Covariance Estimation (pdf, software)
C. Hsieh, I. Dhillon, P. Ravikumar, A. Banerjee.
In Neural Information Processing Systems (NIPS), pp. 2339-2347, December 2012. - The QUIC algorithm for medium-sized problem (dimensionality < 20000):
We implemented the algorithm in C++ and we provide a MEX package and an R package released under the GNU General Public License version 3 or later (GPLv3).
Download the MEX package archive and extract the files. Compile the program using the provided Makefile or use the MEX compiler as follows:
> mex -llapack QUIC.C QUIC-mex.C -output QUIC.[mex|mexa64|mexmaci64|…] - For estimating the inverse covariance matrix when number of random variables > 20000, please use the BigQUIC program. Compile the program using the provided Makefile or use the MEX compiler as follows:
> make