Wiener-Kolmogorov estimators
The estimator of the signal is obtained as
\[\hat{s}_t = k_s \frac{\Psi_s(B)\Psi_s(F)}{\Psi(B)\Psi(F)}y_t=\nu(B, F)y_t\]where $\Psi_i(x)=\frac{\theta_i(x)}{\phi_i(x)\Delta_i(x)}$ and $k_i=\frac{\sigma_i^2}{\sigma^2}$
$\nu(B, F)$ is the Wiener-Kolmogorv filter