#15915: add discrete Gaussian samplers to Sage
-------------------------------------------------+-------------------------
       Reporter:  malb                           |        Owner:
           Type:  enhancement                    |       Status:  new
       Priority:  major                          |    Milestone:  sage-6.4
      Component:  statistics                     |   Resolution:
       Keywords:  sd59                           |    Merged in:
        Authors:  Martin Albrecht                |    Reviewers:  Julian
Report Upstream:  N/A                            |  Rueth
         Branch:                                 |  Work issues:
  caa8ccf49a6ffd2794fd13f0933c47d26336ecbc       |       Commit:
   Dependencies:                                 |     Stopgaps:
-------------------------------------------------+-------------------------

Comment (by nbruin):

 The name `DiscreteGaussianIntegerSampler` might be a bit confusing to some
 people: at first reading it seems you're sampling from the Gaussian
 integers (which are discrete).

 Wouldn't `DiscreteGaussianDistributionSampler` be a more descriptive name?
 I get the impression that Discrete Gaussian Distribution is normally used
 for the distributions on the integers you're referring to.

 The fact that you truncate at `tau*sigma` also seems to disqualify this
 from legitimately being called "sampling from a discrete gaussian
 distribution", but I hope this is standard abuse of terminology in the
 field (and the difference is hard to see).

 It also suggests that if the rejection rate is high (it would be for
 larger values of `tau`) it might be worth sampling from a binomial
 distribution with adapted rejection rates. The shape should be much closer
 to your desired distribution and hence should lead to a much lower
 rejection rate. It depends a bit on whether the new rejection
 probabilities are sufficiently easy to compute to be worth it.

--
Ticket URL: <http://trac.sagemath.org/ticket/15915#comment:56>
Sage <http://www.sagemath.org>
Sage: Creating a Viable Open Source Alternative to Magma, Maple, Mathematica, 
and MATLAB

-- 
You received this message because you are subscribed to the Google Groups 
"sage-trac" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sage-trac.
For more options, visit https://groups.google.com/d/optout.

Reply via email to