“…The priors used for each of the model parameters and hyperparameters (except for ρ) are the same as were used in [18] and are summarized in Table I, where B, G, and U stand for the Bernoulli, gamma, and uniform distributions, respectively, δ is the Dirac delta function, m i denotes the ith row of M , 1 R + (x) = 1 is the indicator function on R + , and α = β = 1. In this table, we can see that each row x i is either jointly zero or nonzero, which is indicated by the discrete variable z i .…”