. Author manuscript; available in PMC: 2019 Oct 1.

Published in final edited form as: J Comput Neurosci. 2018 Sep 6;45(2):83–101. doi: 10.1007/s10827-018-0692-x

Algorithm 4.

GAR - Empirical Bayes

Input:

X_{n}

and W; start values of Ω and Θ.

1. E-step: For i < j, approximate

{\bar{ω}}_{i j} = E [| ω_{i j} | | X_{n}, Θ, W]

2. M-STEP: Iterate a)-b) until convergence:

a) For i = 1,..., d, update α_i according to Eq. (21).

b) Obtain g as the rate function of the Gamma regression of

y_{i j} = 2 α_{i} α_{j} {\bar{ω}}_{i i}

on W_ij, i < j.

3. Iterate 1–2 until convergence.

Output: Estimate of Θ.