Variational Sparse Bayesian Learning for Estimation of Gaussian Mixture Distributed Wireless Channels

. 2021 Sep 28;23(10):1268. doi: 10.3390/e23101268

Algorithm 1: Initialization

Set L = 1, α_{l} = 0, r_{n k} = 0, I = Ø; x_{l} \leftarrow z for l = 1 \to L_{0} Initialize q (θ_{l}) from (25) Initialize q (w_{l}) from (23) Initialize q (α_{l}) from (26) and (27) {\hat{x}}_{l n} \leftarrow y_{n} - \sum_{l = 1}^{L} {\hat{w}}_{l n} s ({\hat{θ}}_{l}) I = I \cup l L = L + 1 end for //remove duplicate estimates for l \in I if \forall d, θ_{d} = θ_{l} w_{l} = w_{l} + w_{d} I = I - d L = L - 1 end if end //initialize GMM parameters for l \in I initialize μ_{l}^{0}, Λ_{l}^{0} from w_{l} by K-means algorithm estimate μ_{l}, Λ_{l}, p (c_{l n k} = 1 | w_{l n}) by EM algorithm r_{l n k} = p (c_{l n k} = 1 | w_{l n}) end