search

LRydin / jumpdiff

JumpDiff: Non-parametric estimator for Jump-diffusion processes for Python
MIT License
45 stars 6 forks source link

Example notation #4

Closed kyrgeorgiou closed 2 years ago

kyrgeorgiou commented 2 years ago

Hi @LRydin, thank you for this great package. I have a question regarding the notation in the definition of the jump term. The parameter xi is referred to as the jump amplitude, which is however normally distributed with variance sigma_xi. Is the xi parameter as set in the example actually the standard deviation of the jump amplitude?

Furthermore, is there a way to set the initial condition, e.g., X_0 = 0?

Thanks again! Kyriakos

LRydin commented 2 years ago

Hey @kyrgeorgiou! Sorry for the delay, I totally missed the notification!

So, the xi, the jump amplitude, should be the variance of the normally distributed jumps. You can see in the code in L.109 that is enters np.random.random() as scale=np.sqrt(xi), and scale expected a standard deviation.

To insert an initial condition to jd_process(), just use init=X_0, with your desired X_0.

Hope this helps!

kyrgeorgiou commented 2 years ago

Hi @LRydin, Great, thanks so much for replying! This indeed helped :)