mutualInfoVAR2: Mutual information analysis of the VAR(2) model

In ragt2ridges: Ridge Estimation of Vector Auto-Regressive (VAR) Processes

Description

Evaluate, within the VAR(2) model, the mutual information between a variate at one time point and the variates at a future time point.

Usage

mutualInfoVAR2(A1, A2, SigmaE, T)

Arguments

A1	A matrix \mathbf{A}_1 of lag one autoregression parameters.
A2	A matrix \mathbf{A}_2 of lag two autoregression parameters.
SigmaE	Covariance matrix of the errors (innovations).
T	Positive integer of length one specifying the lag between time points for which the mutual informations are to be evaluated.

Value

Object of class numeric with elements corresponding to the mutual informations.

Author(s)

Wessel N. van Wieringen <[e-mail]>

References

Miok, V., Wilting, S.M., Van Wieringen, W.N. (2019), “Ridge estimation of network models from time-course omics data”, Biometrical Journal, 61(2), 391-405.

See Also

mutualInfoVAR1, ridgeVAR2.

Examples

# set dimensions (p=covariates, n=individuals, T=time points)
p <- 3; n <- 12; T <- 10

# set model parameters
SigmaE <- diag(p)/4
A1     <- -createA(p, "clique", nCliques=1, nonzeroA=0.1)
A2     <- t(createA(p, "chain", nBands=1, nonzeroA=0.1))

# generate data
Y <- dataVAR2(n, T, A1, A2, SigmaE)

# fit VAR(1) model
VAR2hat <- ridgeVAR2(Y, 1, 1, 1)

# impulse response analysis
mutualInfoVAR2(VAR2hat$A1, VAR2hat$A2, solve(symm(VAR2hat$P)), 10)

ragt2ridges documentation built on Jan. 28, 2020, 5:08 p.m.