Variable Selection and Joint Estimation of Mean and Covariance Models with an Application to eQTL Data

Algorithm 3

Variable selection in the joint estimation.

repeat

Update by (5) and compute with from Algorithm 2.

Update , and using obtained from the previous step.

until the log-likelihood function converges.

while BIC decreases do

Compute Wald statistic of deleting each and in the current model.

Delete one variable (i.e., parameter) with the minimum Wald statistic.

repeat

Update by (5) and compute with from the previous step.

Update , and using obtained from the previous step.

until the log-likelihood function converges.

end while

The optimal model is chosen by BIC.