Joseph K. Bradley, Aapo Kyrola, Danny Bickson, and Carlos Guestrin (2011). "Parallel Coordinate Descent for L1-Regularized Loss Minimization." International Conference on Machine Learning (ICML 2011). http://arxiv.org/abs/1105.5379
1) For running as a mex code called from Matlab Just run:
2) For running as a C application, using MatrixMarket input format:
Current build is only tested on Linux, so you might need to modify the Makefile to suit your system.
We use the following cost function formulation. For Lasso: argmin_x sum_i [(A_ix - y_i)^2 + lambda * |x|_1] For sparse logistic regression: argmin_x sum_i [-log(1 + exp(-y_i * x A_i) ) + lambda * |x|_1]
where |x|_1 is the first norm (sum of absolute value of the vector x).
1) Do not call the mex-library directly. Instead use the provided Matlab-scripts shotgun_logreg.m and shotgun_lasso.m.
Both have same signature:
They return the optimized feature/weight-vector. For tuning the parameters, please modify the scripts. A more user-friendly options-passing will be provided later.
For an example, see example/ directory.
2) RUNNING AS A STANDALONE C PROGRAM: Matrix and vector files are mandaroty inputs
Usage: ./mm_lasso -m matrix A in sparse matrix market format -v vector y in sparse matrix market format -o output file name (will contain solution vector x, default is x.mtx) -a algorithm (1=lasso, 2=logitic regresion, 3 = find min lambda for all zero solution) -t convergence threshold (default 1e-5) -k solution path length (for lasso) -i max_iter (default 100) -n num_threads (default 2) -l lammbda - positive weight constant (default 1) -V verbose: 1=verbose, 0=quiet (default 0)
Provided code is not exactly same as the one we used for running our experiments for the ICML 2011 paper. Particularly this code runs slower sequentially, because special code for running with only one cpu has been removed for clarity. Parallel code needs to do some extra work compared to sequential algorithm, and therefore for fairness we had special versions for the sequential tests.
This version uses OpenMP for parallel execution. For the paper, we used CILK++. Since CILK is not as widely available as OpenMP, we decided to switch for the source code release.
This code is is not well tested and you should not rely on it on any mission-critical tasks.