LINKS OF SOFTWARE / CODES

  • Subset-Selection

Github Repository: [I] NEW Fast algorithms [code in C++ with R-wrapper]: Link

Github Repository: [II] Selection with Shrinkage [Python and Gurobi]: Link

[I] Fast Algorithms based on coordinate descent and combinatorial local search written in C++ with R interface (work in in progress) (Hussein Hazimeh and Rahul Mazumder)

[II] based on the paper: "Subset Selection with Shrinkage: Sparse Linear Modeling when the SNR is low’’, Rahul Mazumder, Peter Radchenko, Antoine Dedieu; 2017. pdf



  • Factor Analysis

Github Repository: Link

based on the paper: "Certifiably Optimal Low Rank Factor Analysis ’’, Dimitris Bertsimas, Martin Copenhaver, Rahul Mazumder; 2017. Journal of Machine Learning Research 18(29):1-53 pdf



  • Trimmed Lasso

Github Repository: Link

based on the paper: ‘‘The Trimmed Lasso: Sparsity and Robustness’’, Dimitris Bertsimas, Martin Copenhaver and Rahul Mazumder; 2017. pdf



  • (Convex) Matrix Completion: “Soft-Impute”

Matlab package: MATLAB-files

R package (CRAN): here

based on the paper : ‘‘Spectral Regularization Algorithms for Learning Large Incomplete Matrices", Rahul Mazumder, Trevor Hastie and Rob Tibshirani, 2010, Journal of Machine Learning Research 11, 2287-2322. pdf



  • (Nonconvex) Matrix Completion: “NC-Impute”

Github Repository: Link



  • Non-Convex penalized regression : “SparseNet” Rpackage

based on the paper : ‘‘SparseNet: Coordinate Descent with Non-Convex Penalties" Rahul Mazumder, Jerome Friedman and Trevor Hastie, 2011, Journal of American Statistical Association, Theory and Methods, 106(495):1125 - 1138.



based on the paper : ‘‘Exact covariance thresholding into connected components for large-scale Graphical Lasso", Rahul Mazumder and Trevor Hastie, 2011, submitted, ArXiv-link; published in JMLR 2012.



  • Graphical Lasso on the Primal (dpglasso) : R-package

based on the paper : “The graphical lasso: New insights and alternatives”, Rahul Mazumder and Trevor Hastie; EJS 2012, Vol(6)



  • Bandwith Extension via Convex Optimization : Matlab Files and demo available from link