Skip to main content
5 events
when toggle format what by license comment
Oct 6, 2013 at 14:42 vote accept Ofer
Oct 6, 2013 at 14:42 comment added Ofer Yes, thank you!!! for some reason I was convinced that Levy measure of the joint process should be $\delta_{0}(y_{1})\frac{\alpha}{\Gamma(1-\alpha)}y_{2}^{-\alpha-1}dy_{2}$.
Oct 6, 2013 at 11:38 comment added Joris Bierkens I would try as compensator for the Poisson random measure $\nu(dy) = \delta_0(d y_1) \nu_2(d y_2) + \nu_1(d y_1) \delta_0(d y_2)$. The drift is essentially $b = (b_1, b_2)$, corrected for the changed unit ball ( $1_{\{|y_1| \leq 1\}} 1_{\{|y_2| \leq 1 \}}$ which becomes $1_{\{ ||(y_1, y_2)|| \leq 1 \}}$). Does this work out for you?
Oct 3, 2013 at 16:43 comment added Ofer Thanks, but how do I find the Levy triplet of $(N_t,D_t)$ - its Levy measure, drift and Gaussian component(which I know is zero since this is a strictly increasing process)?
Oct 3, 2013 at 16:01 history answered Joris Bierkens CC BY-SA 3.0