Skip to main content
. 2017 Jun 14;14(131):20170332. doi: 10.1098/rsif.2017.0332

Figure 2.

Figure 2.

Representation of the proposed Bayesian dynamic elastic-net approach as a probabilistic graphical model. The hidden influences wl form a Markov chain over all time points l = 1, …, T and are directly dependent on the shared parameters Inline graphic and λ2. Since the outcome of one integration step represents the initial value for the next integration step, the system state variables Inline graphic are also successively dependent.