No packages match

huge - High-Dimensional Undirected Graph Estimation

Provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation, the graphical lasso, or the TIGER (tuning-insensitive graph estimation and regression) method, and the first two can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.

Last updated

openblascppopenmp

10.18 score 10 dependents 824 scripts 8.4k downloads

flare - Family of Lasso Regression

Provides implementations of a family of Lasso variants, including Dantzig Selector, LAD Lasso, SQRT Lasso, and Lq Lasso, for estimating high-dimensional sparse linear models. We adopt the alternating direction method of multipliers and convert the original optimization problem into a sequence of L1-penalized least-squares minimization problems that are efficiently solved by linearization and multi-stage screening. In addition to sparse linear model estimation, we provide extensions of these methods to sparse Gaussian graphical model estimation, including TIGER and CLIME, using either L1 or adaptive penalties. Missing values can be tolerated for Dantzig selector and CLIME. Computation is memory-optimized using sparse matrix output. For more information, see <https://www.jmlr.org/papers/volume16/li15a/li15a.pdf>.

Last updated

5.11 score 1 stars 4 dependents 213 scripts 963 downloads

picasso - Sparse Learning with Convex and Concave Penalties

Fast tools for fitting sparse generalized linear models with convex penalties (lasso) and concave penalties (smoothly clipped absolute deviation and minimax concave penalty). Computation uses multi-stage convex relaxation and pathwise coordinate optimization with warm starts, active-set updates, and screening rules. Core solvers are implemented in C++, and coefficient paths are stored as sparse matrices for memory efficiency.

Last updated

cpp

2.60 score 1 stars 40 scripts 297 downloads