Graphical VAR for Experience Sampling Data

Estimates within and between time point interactions in experience sampling data, using the Graphical vector autoregression model in combination with regularization. See also Epskamp, Waldorp, Mottus & Borsboom (2018) .


Changes in Version 0.2.2: o Fixed a CRAN warning o Added 'lags' and 'likelihood' argments to graphicalVAR o Fixed message about deprecated functions

Changes in Version 0.2.1: o Data is now stored in the output of graphicalVAR o Fixed a bug leading to NA in any column to delete a row o Added 'likelihood' argument to mimic sparseTSCGM 2.5 behavior

Changes in Version 0.2: o Added 'tsData' function to prepare data o graphicalVAR now supports multiple subjects (fixed effects only) and day effects o 'beepvar' can now be used to handle missing beeps o Added 'mimic' argument to 'graphicalVAR' to mimic 0.1.2 and 0.1.4 behavior. o Lambda sequence for kappa now uses the maximum absolute correlation as maximal value rather than 1. o Added 'mlGraphicalVAR' for pooled and individual estimation of N > 1 datasets o Added 'simMLgvar' to simulate N > 1 data o Fixed a bug in 'graphicalVARsim' where mean structure was not used in data generation

Changes in Version 0.1.4: o 'graphicalVAR' again standardizes variables before running by default. Can be controlled using the 'scale' argument o New arguments to 'graphicalVAR': o deleteMissings o penalize.diagonal o lambda_min_kappa o lambda_min_beta o scale o Added 'randomGVARmodel' function to simulate graphicalVAR model matrices o Added unregularized estimation when both lambda_kappa = 0 and lambda_beta = 0 o Greatly updated the tuning parameter sequence generating algorithm. The sequence should now be better chosen. However, note that this change leads to different estimated networks as with previous graphicalVAR versions (as different LASSO tuning parameters are used) o Added 'lbound' and 'ubound' arguments to graphicalVARsim

Changes in Version 0.1.3: o graphicalVAR now only centers and does not standardize

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.3 by Sacha Epskamp, 3 months ago

Browse source code at

Authors: Sacha Epskamp [aut, cre] , Eren Asena [ctb]

Documentation:   PDF Manual  

Task views: Time Series Analysis, Psychometric Models and Methods

GPL (>= 2) license

Imports Rcpp, Matrix, glasso, glmnet, mvtnorm, qgraph, dplyr, methods, igraph, rlang

Linking to Rcpp, RcppArmadillo

Imported by mlVAR.

Suggested by bootnet, psychonetrics.

See at CRAN