The Bayesian model follows the standard NSSM defined in Equation (4.13):

The basic structure of the model can be seen in Figure 5.1.

The nonlinear functions and of the NSSM are modelled by MLP networks. The networks have one hidden layer and can thus be written as in Equation (4.15):

where and are the weight matrices of the networks and and are the bias vectors.

Because the data are assumed to be generated by a continuous dynamical system, it is reasonable to assume that the values of the hidden states do not change very much from one time index to the next one. Therefore the MLP network representing the function is used to model only the change.