SPINCOVSE version 0 -- a MATLAB software for sparse inverse covariance selection based on inexact interior-point methods

Kim-Chuan Toh and Lu Li

The software was first released in 2011. The software is designed to solve sparse inverse covariance selection problems of the form:

\( \qquad \min_X \big\{ {\rm Trace}(S X) - {\rm logdet}(X) + \sum_{(i,j)\not\in E} H_{ij} |X_{ij}| \mid X_{ij} = 0 \;\forall\; (i,j) \in E,\, X \succeq 0 \big\} \)

where \(S\) is the given data, \(H_{ij}\) are given positive weights, and \(E\) is a given set of index pairs.
Important note: this is a research software. It is not intended nor designed to be a general purpose software at the moment.
Citation:
  • Lu Li and Kim-Chuan Toh An inexact interior point method for L1-regularized sparse covariance selection, Mathematical Programming Computation, 2 (2010), pp. 291--315.