Thanks, I see. So there is a sample generator for Levy. It might be not that easy for stable distributions. I should start with naive implementation.
Pavel On Dec 21, 2010, at 23:32 PM, Phil Steitz wrote: > On Tue, Dec 21, 2010 at 5:03 PM, Pavel Ryzhov <[email protected]>wrote: > >> No, right now. >> >> But i'll definitely make sample generation as i'm writing small Monte Carlo >> engine. >> My current plan is: >> 1. Levy distribution as the first step for contribution. >> 2. Stable distribution and generic distributions by characteristic >> function. I've already started working on it. >> 3. Generators for 1 and 2. >> 4. Generator for Ito process. I'm not sure about Stratanovich processes as >> they are rarely used in finance. >> > > Have a look at o.a.c.m.distribution if you have not already done so. If you > just provide an implementation of the distribution, a default > inversion-based sampler will be inherited from the base class. If you want > to improve the sampling, you can (eventually) override the sample() method. > > Phil > > >> >> Regards, >> Pavel >> >> On Dec 21, 2010, at 22:27 PM, Ted Dunning wrote: >> >>> Sounds interesting to have to me. >>> >>> Do you generate samples? >>> >>> On Tue, Dec 21, 2010 at 12:38 PM, Pavel Ryzhov <[email protected] >>> wrote: >>> >>>> Hi, >>>> >>>> I've implemented Levy distribution on top of commons-math. The >>>> implementation is pretty straightforward by >>>> http://en.wikipedia.org/wiki/Lévy_distribution<http://en.wikipedia.org/wiki/L%C3%A9vy_distribution> >> <http://en.wikipedia.org/wiki/L%C3%A9vy_distribution>So it was not a big >> deal. The distribution is of interest to the financial >>>> modeling, so it might worth to include it into the library. >>>> >>>> So, do we need it in commons-math? >>>> >>>> Regards, >>>> Pavel >>>> >>>> >>>> --------------------------------------------------------------------- >>>> To unsubscribe, e-mail: [email protected] >>>> For additional commands, e-mail: [email protected] >>>> >>>> >> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [email protected] >> For additional commands, e-mail: [email protected] >> >> --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
