Learn R Programming

mglasso (version 0.1.2)

Multiscale Graphical Lasso

Description

Inference of Multiscale graphical models with neighborhood selection approach. The method is based on solving a convex optimization problem combining a Lasso and fused-group Lasso penalties. This allows to infer simultaneously a conditional independence graph and a clustering partition. The optimization is based on the Continuation with Nesterov smoothing in a Shrinkage-Thresholding Algorithm solver (Hadj-Selem et al. 2018) implemented in python.

Copy Link

Version

Install

install.packages('mglasso')

Monthly Downloads

223

Version

0.1.2

License

MIT + file LICENSE

Maintainer

Edmond Sanou

Last Published

September 8th, 2022

Functions in mglasso (0.1.2)

symmetrize

Apply symmetrization on estimated graph
fun_lines

weighted sum/difference of two regression vectors
adj_mat

Adjacency matrix
beta_to_vector

Transform a matrix of regression coefficients to vector removing the diagonal
beta_ols

Initialize regression matrix
cost

Mglasso cost function
conesta

CONESTA solver.
dist_beta

Compute distance matrix between regression vectors
plot_mglasso

Plot mglasso function output.
image_sparse

Plot the image of a matrix
install_pylearn_parsimony

Install the python library pylearn-parsimony and other required libraries
precision_to_regression

Compute precision matrix from regression vectors
mglasso

Inference of Multiscale Gaussian Graphical Model.
merge_clusters

compute clusters partition from pairs of variables to merge
plot_clusterpath

Plot MGLasso Clusterpath