Description Usage Arguments Value Note Author(s) References See Also Examples
View source: R/loglikLOOCVVAR1fused.r
Evaluation of the (minus) leave-one-out cross-validated log-likelihood of multiple VAR(1) models for given choices of the (fused) ridge penalty parameters (λ_a, λ_f and λ_{ω} for the autoregression coefficient matrice \mathbf{A}_g's and the inverse error covariance matrix \mathbf{Ω}_{\varepsilon} (=\mathbf{Σ_{\varepsilon}^{-1}}), respectively). The functions also works with a (possibly) unbalanced experimental set-up. The VAR(1)-processes are assumed to have mean zero.
1 | loglikLOOCVVAR1fused(lambdas, Y, id, unbalanced=matrix(nrow=0, ncol=2), ...)
|
lambdas |
A |
Y |
Three-dimensional |
id |
A vector with group indices comprising of integers only. First group is represented by '0', the next by '1', and so on until the last. |
unbalanced |
A |
... |
Other arguments to be passed to |
A numeric
of length one: the minus (!) LOOCV log-likelihood.
The minus LOOCV log-likelihood is returned as standard optimization procedures in R like nlminb
and constrOptim
minimize (rather then maximize). Hence, by providing the minus LOOCV log-likelihood the function loglikLOOCVVAR1fused
can directly used by these optimization procedures.
Wessel N. van Wieringen <w.vanwieringen@vumc.nl>
Miok, V., Wilting, S.M., Van Wieringen, W.N. (2018), “Ridge estimation of network models from time-course omics data”, Biometrical Journal, <DOI:10.1002/bimj.201700195>.
ridgeP
and ridgeVAR1fused
.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 | # set dimensions (p=covariates, n=individuals, T=time points, G=groups)
p <- 3; n <- 12; T <- 10; G <- 3
# set model parameters
SigmaE <- matrix(1/2, p, p)
diag(SigmaE) <- 1
A1 <- -createA(p, "clique", nCliques=1, nonzeroA=0.1)
A2 <- t(createA(p, "chain", nBands=1, nonzeroA=0.1))
A3 <- (A1 + A2) / 2
# generate data
Y1 <- dataVAR1(n/G, T, A1, SigmaE)
Y2 <- dataVAR1(n/G, T, A2, SigmaE)
Y3 <- dataVAR1(n/G, T, A3, SigmaE)
Y <- abind::abind(Y1, Y2, Y3, along=3)
id <- c(rep(1, n/G), rep(2, n/G), rep(3, n/G))-1
## determine optimal values of the penalty parameters
## Not run: optLambdas <- constrOptim(c(1,1,1), loglikLOOCVVAR1fused, gr=NULL,
## Not run: ui=diag(3), ci=c(0,0,0), Y=Y, id=id
## Not run: control=list(reltol=0.01))$par
# ridge ML estimation of the VAR(1) parameter estimates with
# optimal penalty parameters
optLambdas <- c(0.1, 0.1, 0.1)
VAR1hats <- ridgeVAR1fused(Y, id, optLambdas[1], optLambdas[2], optLambdas[3])
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.