Bayesian Sparse Estimation Using Double Lomax Priors

Figure 6

The update traces of latent variables and noise parameter when both ARD and DL converge to the global minima (shown in Figure 5(b)).

(a) log⟨𝛼_𝑚⁻¹⟩ in ARD

(b) log⟨𝛼_𝑚⁻¹⟩ in DL

(c) log⟨𝜐²_𝑚⟩ in DL

(d) ⟨𝛼_𝑚⁻¹⟩−⟨𝜐²_𝑚⟩

(e) ⟨𝛼_𝑚⁻¹⟩−⟨𝜐²_𝑚⟩

(f) ⟨𝛽⟩^−0.5