Talk:Convolution random number generator
![]() | This article is rated Stub-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects: | |||||||||||||||||||||||||||
|
this isnt a monte carlo method. i have undone your categorization and put it back into the sampling techniques category. —Preceding unsigned comment added by Marvinklein (talk • contribs)
clarification
[edit]represents a random variable sampled from the Erlang distribution. Convolution sampling is a method which, for example, lets you sample the erlang distribution using only random variables from the exponential distribution. —Preceding unsigned comment added by Marvinklein (talk • contribs)
What does this mean?
[edit]I'm confused by this article. Specifically, I'm not sure what the operator is supposed to represent.
I searched Google Scholar for "convolution sampling". Most of the hits I got used constructions like "convolution, sampling", or "... convolution. Sampling ..." – in other words, I couldn't find a reliable reference that used "convolution sampling" as a specific concept. I did hit the German-English dictionary to understand "Erlang", and that seems to occur only in words that mean something like "to get", or "to obtain".
Can somebody shed some light on this one? DavidCBryant 16:25, 18 April 2007 (UTC)
Meaning of this article
[edit]It looks as if the point of this article is that some probability distributions can be realized as convolutions, so that to generate a random variable from such a distribution, you first generate independent r.v.s from certain other distributions and then simply add them.
Is that the point of the article? Michael Hardy (talk) 12:24, 20 October 2009 (UTC)