Re: anyone want to implement these transformation functions?

[Jason Stover]

I'd be happy to implement these, but I don't know when I can get to
them. I have written the guts of some of these functions, and those
guts are attached, if anyone else wants to put them in PSPP before
I can.

I wrote these functions for GSL, though the maintainer hasn't put them
in. (I don't know if he will.) Gutting them wouldn't be difficult.  I
ran some accuracy tests on them, which I can send, if you like.

Attached is code for the binomial cdf, the geometric cdf, the hypergeometric 
cdf,
the Poisson cdf. Some of these require the beta cdf, whose code is
also attached.

-Jason

On Sun, Mar 06, 2005 at 10:30:54PM -0800, Ben Pfaff wrote:
> We don't have implementations of the following functions for use
> in PSPP expressions:
> 
>     function NCDF.CHISQ (x >= 0, df > 0, c) = unimplemented;
>     function NPDF.CHISQ (x >= 0, df > 0, c) = unimplemented;
>     function NCDF.F (x >= 0, df1 > 0, df2 > 0, lambda >= 0) = unimplemented;
>     function NPDF.F (x >= 0, df1 > 0, df2 > 0, lmabda >= 0) = unimplemented;
>     function SIG.F (x >= 0, df1 > 0, df2 > 0) = unimplemented;
>     function CDF.HALFNRM (x, a, b > 0) = unimplemented;
>     function IDF.HALFNRM (P > 0 && P < 1, a, b > 0) = unimplemented;
>     function PDF.HALFNRM (x, a, b > 0) = unimplemented;
>     no_opt function RV.HALFNRM (a, b > 0) = unimplemented;
>     function CDF.IGAUSS (x > 0, a > 0, b > 0) = unimplemented;
>     function IDF.IGAUSS (P >= 0 && P < 1, a > 0, b > 0) = unimplemented;
>     function PDF.IGAUSS (x > 0, a > 0, b > 0) = unimplemented;
>     no_opt function RV.IGAUSS (a > 0, b > 0) = unimplemented;
>     function CDF.SMOD (x > 0, a >= 1, b >= 1) = unimplemented;
>     function IDF.SMOD (P >= 0 && P < 1, a >= 1, b >= 1) = unimplemented;
>     function CDF.SRANGE (x > 0, a >= 1, b >= 1) = unimplemented;
>     function IDF.SRANGE (P >= 0 && P < 1, a >= 1, b >= 1) = unimplemented;
>     function NCDF.T (x, df > 0, nc) = unimplemented;
>     function NPDF.T (x, df > 0, nc) = unimplemented;
>     function CDF.BINOM (k, n > 0 && n == floor (n), p >= 0 && p <= 1)
>          = unimplemented;
>     function CDF.GEOM (k >= 1 && k == floor (k), p >= 0 && p <= 1) = 
> unimplemented;
>     function CDF.HYPER (k >= 0 && k == floor (k) && k <= c,
>                         a > 0 && a == floor (a),
>                         b > 0 && b == floor (b) && b <= a,
>                         c > 0 && c == floor (c) && c <= a)
>          = unimplemented;
>     function CDF.NEGBIN (k >= 1, n == floor (n), p > 0 && p <= 1) = 
> unimplemented;
>     function CDF.POISSON (k >= 0 && k == floor (k), mu > 0) = unimplemented;
> 
> (The above is in the syntax used for operations.def, which is
> terse but expressive.)
> 
> Anyone want to implement any of them?  As far as I can tell,
> these functions are not part of GSL.
> -- 
> "Platonically Evil Monkey has been symbolically representing the darkest 
>  fears of humanity since the dawn of literature and religion, and I think
>  I speak for everyone when I give it a sidelong glance of uneasy recognition 
>  this evening." --Scrymarch
> 
> 
> _______________________________________________
> pspp-dev mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/pspp-dev

-- 
address@hidden
SDF Public Access UNIX System - http://sdf.lonestar.org

hypergeometric.c
Description: Text document

negbinom.c
Description: Text document

binomial.c
Description: Text document

poisson.c
Description: Text document

geometric.c
Description: Text document

betadistinv.c
Description: Text document