graphicalVAR (version 0.3.4)

mlGraphicalVAR: Pooled and individual graphical VAR estimation

Description

This function fits fixed effect temporal and contemporaneous networks over multiple subjects and runs separate graphical VAR models per subject. The algorithm does: (1) pool all data, within-subject center variables and run graphicalVAR to obtain fixed effects, (2) run EBICglasso on subject means to obtain a between-subjects network, (3) run graphicalVAR on data of every subject to obtain individual networks. See arxiv.org/abs/1609.04156 for more details.

Usage

mlGraphicalVAR(data, vars, beepvar, dayvar, idvar, scale = TRUE,
              centerWithin = TRUE, gamma = 0.5, verbose = TRUE,
              subjectNetworks = TRUE, lambda_min_kappa_fixed = 0.001,
              lambda_min_beta_fixed = 0.001, lambda_min_kappa = 0.05,
              lambda_min_beta = lambda_min_kappa, lambda_min_glasso = 0.01,
              ...)

Value

A "mlGraphicalVAR" object with the following elements:

fixedPCC

Estimated fixed effects (partial contemporaneous correlations) of contemporaneous effects

fixedPDC

Estimated fixed effects (partial directed correlations) of temporal effects

fixedResults

Full object of pooled data estimation (fixed effects)

betweenNet

Estimated between-subjects network (partial correlations)

ids

Vector of subject IDs

subjectPCC

List of estimated individual contemporaneous networks

subjectPDC

List of estimated individual directed networks

subjecResults

List of full results of individual estimations

Arguments

data

Data frame

vars

Vectors of variables to include in the analysis

beepvar

String indicating assessment beep per day (if missing, is added). Adding this argument will cause non-consecutive beeps to be treated as missing!

dayvar

String indicating assessment day. Adding this argument makes sure that the first measurement of a day is not regressed on the last measurement of the previous day. IMPORTANT: only add this if the data has multiple observations per day.

idvar

String indicating the subject ID

scale

Logical, should variables be standardized before estimation?

centerWithin

Logical, should subject data be within-person centered before estimating fixed effects?

gamma

EBIC tuning parameter.

verbose

Logical indicating if console messages and the progress bar should be shown.

subjectNetworks

TRUE to estimate all subject numbers, or a vector with IDs of which subject numbers should be estimated.

lambda_min_kappa_fixed

Multiplier of maximal tuning parameter

lambda_min_beta_fixed

Multiplier of maximal tuning parameter

lambda_min_kappa

Multiplier of maximal tuning parameter

lambda_min_beta

Multiplier of maximal tuning parameter

lambda_min_glasso

Multiplier of maximal tuning parameter

...

Arguments sent to graphicalVAR

Author

Sacha Epskamp <[e-mail]>

References

Epskamp, S., Waldorp, L. J., Mottus, R., & Borsboom, D. Discovering Psychological Dynamics: The Gaussian Graphical Model in Cross-sectional and Time-series Data.

See Also

graphicalVAR

Examples

if (FALSE) {
# Simulate data:
Sim <- simMLgvar(nTime = 50, nPerson = 20, nVar = 3)

# Estimate model:
Res <- mlGraphicalVAR(Sim$data, vars = Sim$vars, idvar = Sim$idvar)

layout(t(1:2))
library("qgraph")

# Temporal fixed effects
qgraph(Res$fixedPDC, title = "Estimated fixed PDC", layout = "circle")
qgraph(Sim$fixedPDC, title = "Simulated fixed PDC", layout = "circle")

# Contemporaneous fixed effects
qgraph(Res$fixedPCC, title = "Estimated fixed PCC", layout = "circle")
qgraph(Sim$fixedPCC, title = "Simulated fixed PCC", layout = "circle")
}