Hello. Le mar. 1 déc. 2020 à 01:42, Marko Malenic <mmalen...@gmail.com> a écrit : > > Hi, > > I'm a bit new to all this stuff, so bear with me while I ask some questions > :) > > There's a few ways to do this. > > In terms of number generation, there's a few algorithms, some of which at > described at: > https://en.wikipedia.org/wiki/Truncated_normal_distribution#Computational_methods > Any preferences on how to generate the numbers? > > I noticed sampling is split off to commons rng. > Should another sampler be added, depending on the algorithm? > Or maybe just using inverse transform sampling would be okay.
Some of the implemented distributions use the inverse transform. It's fine and sane to not do everything at once. ;-) Indeed, if you implement another sampler, it must go into the "sampling" module of "Commons RNG", reusing functionality already implemented there (if applicable). Regards, Gilles > [...] --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
