covres: Sample Covariance Matrix of Residuals
In Tlasso: Non-Convex Optimization and Statistical Inference for Sparse Tensor Graphical Models

Description Usage Arguments Details Value Author(s) See Also Examples

View source: R/covres.R

Generate sample covariance matrix of residuals (includes diagnoal) described in Sun et al. (2016).

1	covres(data, Omega.list, k = 1)

`data`	tensor object stored in a m1 * m2 * ... * mK * n array, where n is sample size and mk is dimension of the kth tensor mode.
`Omega.list`	list of precision matrices of tensor, i.e., `Omega.list[[k]]` is precision matrix for the kth tensor mode, 1 <= k <= K .
`k`	index of interested mode, default is 1.

This function computes sample covariance of residuals and is the basis for support recovery procedure in Sun et al. (2016). Note that output matrix includes diagnoal while bias corrected matrix (output of biascor) for inference is off-diagnoal, see Sun et al. (2016) for details. Elements in Omega.list are true precision matrices or estimation of the true ones, the latter can be output of Tlasso.fit.

A matrix whose (i,j) entry (includes diagnoal) is sample covariance of the ith and jth residuals in the kth mode. See Sun et al. (2016) for details.

Will Wei Sun, Zhaoran Wang, Xiang Lyu, Han Liu, Guang Cheng.

varcor, biascor

m.vec = c(5,5,5)  # dimensionality of a tensor 
n = 5   # sample size 
k=1 # index of interested mode
lambda.thm = 20*c( sqrt(log(m.vec[1])/(n*prod(m.vec))), 
                   sqrt(log(m.vec[2])/(n*prod(m.vec))), 
                   sqrt(log(m.vec[3])/(n*prod(m.vec))))
DATA=Trnorm(n,m.vec,type='Chain') 
# obersavations from tensor normal distribution
out.tlasso = Tlasso.fit(DATA,T=1,lambda.vec = lambda.thm)   
# output is a list of estimation of precision matrices
rho=covres(DATA, out.tlasso, k = k) # sample covariance of residuals, including diagnoal 
rho