A Pathwise Algorithm for Covariance Selection

  • TITLE: A Pathwise Algorithm for Covariance Selection.

  • AUTHORS: Vijay Krishnamurthy, Selin Damla Ahipasaoglu, Alexandre d'Aspremont

  • ABSTRACT: Covariance selection seeks to estimate a covariance matrix by maximum likelihood while restricting the number of nonzero inverse covariance matrix coefficients. A single penalty parameter usually controls the tradeoff between log likelihood and sparsity in the inverse matrix. We describe an efficient algorithm for computing a full regularization path of solutions to this problem.

  • STATUS: Optimization for Machine Learning, S. Sra, S. Nowozin and S. Wright, editors, MIT Press, 2011.

  • ArXiv PREPRINT: 0908.0143

  • PAPER: A Pathwise Algorithm for Covariance Selection in pdf.

  • CODE: The COVPATH library is available on CRAN