Tsphere (version 1.0)

covTranspose11: Covariance Estimation.

Description

Inverse row and column covariance estimation for the L1 penalized matrix-variate normal model.

Usage

covTranspose11(xc, rhor, rhoc, row = TRUE, sigi.init = NULL, 
delti.init = NULL, thr = 1e-04, maxit = 1000, trace = TRUE,  thr.glasso
= 1e-04,  maxit.glasso = 1000, pen.diag = TRUE)

Arguments

xc

Centered data matrix.

rhor

Row regularization parameter.

rhoc

Column regularization parameter.

row

Logical. TRUE = Start with row covariance.

sigi.init

Initialization for the row precision matrix. (Optional).

delti.init

Initialization for the column precision matrix. (Optional).

thr

Convergence threshold.

maxit

Maximum number of iterations.

trace

Prints matrix-variate log-likelihood for each iteration.

thr.glasso

Convergence threshold for the graphical lasso.

maxit.glasso

Maximum number of iterations for the graphical lasso.

pen.diag

Logical. Indicates whether the diagonal should be penalized.

Value

• Sigmahat
  Estimated row covariance.
• Deltahat
  Estimated column covariance.
• Sigmaihat
  Estimated sparse row precision matrix.
• Deltaihat
  Estimated sparse column precision matrix.
• loglike
  Trace of the penalized log-likelihood at each iteration.

Details

Estimates row and column precision matrix via L1 penalized Transposable Regularized Covariance Models.

References

G. I. Allen and R. Tibshirani, "Transposable regularized covariance models with an application to missing data imputation", Annals of Applied Statistics, 4:2, 764-790, 2010.

See Also

TransSphere