[prev in list] [next in list] [prev in thread] [next in thread] 

List:       boost-docs
Subject:    Re: [Boost-docs] INdexing
From:       Andrew Sutton <asutton () cs ! kent ! edu>
Date:       2007-09-22 0:11:24
Message-ID: C024C53A-B907-49AF-9811-0E95BD1573AF () cs ! kent ! edu
[Download RAW message or body]

> I don't suppose it's possible to implement as something like:
>
> template <class Distribution>
> class skewed_distribution
> {
>    skewed_distribution(const Distribution&, typename
> Distribution::value_type skewness);
> };
>
> So we get a skewed t-distribution and whatever else for free?

I'm not sure. My gut feeling is no. I think that skewing a  
distribution essentially fundamentally modifies the formulas that,  
for example, define its pdf, and that these modifications differ  
fundamentally from distribution to distribution. These changes would  
probably propagate inconsistently to the moment generating functions.

On the other hand, I may very well be completely wrong :) I'm not  
that great of a mathematician, but it's an interesting question.

Andrew Sutton
asutton@cs.kent.edu



_______________________________________________
Boost-docs mailing list
Boost-docs@lists.boost.org
http://lists.boost.org/mailman/listinfo.cgi/boost-docs
[prev in list] [next in list] [prev in thread] [next in thread]