simulation

number of features (number of nodes)

a vector indicating number of samples and tasks, for example c(100,200,300) for 3 tasks and 100,200 and 300 samples for task 1, 2 and 3

seedNum

positive number that controls sparsity of the generated graphs

positive number that controls sparsity of the shared part of generated graphs

simulate multiple sparse graphs and generate samples

An end-to-end package for learning multiple sparse Gaussian graphical models and nonparanormal models from Heterogeneous Data with Additional Knowledge. It is able to simulate multiple related graphs as well as produce samples drawn from them. Multiple state-of-the-art sparse Gaussian graphical model estimators are included to both multiple and difference estimation. Graph visualization is available in 2D as well as 3D, designed specifically for brain. Moreover, a set of evaluation metrics are integrated for easy exploration with model validity. Finally, classification using graphical model is achieved with Quadratic Discriminant Analysis. The package comes with multiple demos with datasets from various fields. Methods references: SIMULE (Wang B et al. (2017) <doi:10.1007/s10994-017-5635-7>), WSIMULE (Singh C et al. (2017) <arXiv:1709.04090v2>), DIFFEE (Wang B et al. (2018) <arXiv:1710.11223>), JEEK (Wang B et al. (2018) <arXiv:1806.00548>), JGL(Danaher P et al. (2012) <arXiv:1111.0324>) and kdiffnet (Sekhon A et al, preprint for publication).

simulation: simulate multiple sparse graphs and generate samples

Description

Usage

Arguments

Value

Examples