Tune_SSLasso: Tuning SSLasso

In garyganuiuc/SSLasso: Sparse Precision Matrix Estimation with Bayesian Regularization

Usage

Tune_SSLasso(v0, tau, S, n, p_n, p)

Arguments

v0	Tuning parameter v0
tau	Tuning parameter tau, control on the diagonal
S	Sample covariance matrix
n	Sample size
p_n	Matrix dimension
p	The hyperparameter eta

Examples

##---- Should be DIRECTLY executable !! ----
##-- ==>  Define data, use random,
##--	or do  help(data=index)  for the standard data sets.

## The function is currently defined as
function (v0, tau, S, n, p_n, p) 
{
    pb <- txtProgressBar(min = 0, max = length(v0) * 10 * length(tau), 
        style = 3)
    bic = NULL
    for (m in 1:length(p)) {
        for (i in 1:length(v0)) {
            v1 = seq(v0[i] + 0.1, 5 * v0[i], 0.5 * v0[i])
            for (j in 1:length(v1)) {
                for (k in 1:length(tau)) {
                  w = 1
                  l = 1
                  maxiter = 30
                  result1 <- EM_lasso(S, n, p_n, v0[i], v1[j], 
                    maxiter, p[m], tau[k])
                  bic = rbind(bic, list(v0 = v0[i], v1 = v1[j], 
                    tau = tau[k], p = p[m], BIC = BIC_SSLasso(result1$Theta, 
                      S, result1$P, n)))
                  setTxtProgressBar(pb, (m - 1) * length(v0) * 
                    length(v1) * length(tau) + (i - 1) * length(v1) * 
                    length(tau) + (j - 1) * length(tau) + k)
                }
            }
        }
    }
    close(pb)
    return(bic[which.min(bic[, 5]), ])
  }

garyganuiuc/SSLasso documentation built on May 16, 2019, 5:43 p.m.