Research Article

Inverse Dispersion for an Unknown Number of Sources: Model Selection and Uncertainty Analysis

Figure 1

Four iterations of the algorithm of Section 2 showing the samples of source model 𝑚 𝑁 drawn from the posterior distribution 𝑝 ( 𝜃 𝑚 𝑁 , 𝒟 , 𝐼 ) and the computation of 𝐾 𝜖 1 0 for (a) 𝑁 = 1 , l o g ( 𝐾 𝜖 1 0 ) = 6 2 . 4 ± 0 . 2 ; (b) 𝑁 = 2 , l o g ( 𝐾 𝜖 1 0 ) = 5 7 . 3 ± 0 . 3 ; (c) 𝑁 = 3 , l o g ( 𝐾 𝜖 1 0 ) = 3 9 . 1 ± 0 . 3 ; (d) 𝑁 = 4 , l o g ( 𝐾 𝜖 1 0 ) = 0 . 9 2 ± 0 . 3 3 . A solid square denotes a detector.
465320.fig.001a
(a) N = 1 
465320.fig.001b
(b) N = 2
465320.fig.001c
(c) N = 3
465320.fig.001d
(d) N = 4