Figure 8From: Constructing non-stationary Dynamic Bayesian Networks with a flexible lag choosing mechanismComparison of three methods on Arabidopsis T28 data. Left: The posterior probabilities of the numbers of segments (top: FLnsDBNs (λ m = 14, λ s = 2); middle: RJnsDBNs (λ m = 0.005, λ s = 2); bottom: ASnsDBNs). Right: The posterior probabilities of the change points (FLnsDBNs: black solid line; RnsDBNs: magenta dash-dot line; ASnsDBNs: blue dashed line).Back to article page