next up previous
Next: Maximum likelihood estimation Up: OU-processes Previous: Momenta

OU-process in VAR form

For equidistant observation times $t_i=i\,\delta$ the OU-process takes the VAR(1) form
\begin{displaymath}
\xi_{i+1} -\mu_\infty =
e^{\delta\,A}(\xi_i-\mu_\infty) + V_\delta\,\epsilon_i
\end{displaymath} (73)

where the $\epsilon_i$ are iid $\mathcal N(0,1)$.



Markus Mayer 2009-06-22