Prev: Accepting online payment
Next: Question about LLA: how to access a lisp-array
From: allchemist on 14 Apr 2010 12:16
What is the best math-statistics suite for lisp?
"Best" means a good combination of power, efficiency and also being in
actual state (the author didn't forget about it). Operating with
native lisp arrays is very desired.

GSLL provides good statistics support, but there are some problems
with contacting to native arrays.
From: Captain Obvious on 14 Apr 2010 12:43
a> What is the best math-statistics suite for lisp?
a> "Best" means a good combination of power, efficiency and also being in
a> actual state (the author didn't forget about it). Operating with
a> native lisp arrays is very desired.

I don't think this library:

http://common-lisp.net/project/sapaclisp/

meets all your criteria, but it has some good, very well documented math and
statistics related functions, and it uses native lisp data structures.

From: Tamas K Papp on 14 Apr 2010 14:24
On Wed, 14 Apr 2010 09:16:04 -0700, allchemist wrote:

> What is the best math-statistics suite for lisp? "Best" means a good
> combination of power, efficiency and also being in actual state (the
> author didn't forget about it). Operating with native lisp arrays is
> very desired.

What kind of statistics? What do you want to do? You could be more
specific.

I don't think that you are going to find libraries operating on native
arrays that are efficient. The reason for this is that the CL
standard doesn't require the existence of arrays specialized to the
commonly used float types. Eg in ECL,

(upgraded-array-element-type '(complex double-float))

is T. In CLISP,

(upgraded-array-element-type 'double-float)

is T. So either you use general arrays (in which case you won't be
able to get fast code and optimize), or you wrap the arrays. LLA is
the thinnest wrapper I have seen, but it is still experimental. AFAIK
CLEM also does this, but it doesn't support a whole lot of operations
and is written in pure CL.

> GSLL provides good statistics support, but there are some problems with
> contacting to native arrays.

What problems? Last time I checked, it was very easy to convert to
native arrays. If you find that it is not working, contact the GSLL
mailing list.

Tamas
From: allchemist on 14 Apr 2010 14:57
On 14 апр, 22:24, Tamas K Papp <tkp...(a)gmail.com> wrote:
> On Wed, 14 Apr 2010 09:16:04 -0700, allchemist wrote:
> > What is the best math-statistics suite for lisp? "Best" means a good
> > combination of power, efficiency and also being in actual state (the
> > author didn't forget about it). Operating with native lisp arrays is
> > very desired.
>
> What kind of statistics?  What do you want to do?  You could be more
> specific.
Generally saying, statistical learning and statistical pattern
recognition. These leads to a lot of funcalls, getting covariance,
correlation, autocorrelation matrices, also decorrelation routines
(btw, whitening could be done through svd).

> So either you use general arrays (in which case you won't be
> able to get fast code and optimize), or you wrap the arrays.  LLA is
> the thinnest wrapper I have seen, but it is still experimental.  AFAIK
> CLEM also does this, but it doesn't support a whole lot of operations
> and is written in pure CL.
I understand, that the only way is wrapping.

> What problems?  Last time I checked, it was very easy to convert to
> native arrays.  If you find that it is not working, contact the GSLL
> mailing list.

Hm, I didn't manage to do it. Maybe a have too wrong hands.
From: AJ Rossini on 14 Apr 2010 14:57
On Apr 14, 5:16 pm, allchemist <hohlovi...(a)gmail.com> wrote:
> What is the best math-statistics suite for lisp?
> "Best" means a good combination of power, efficiency and also being in
> actual state (the author didn't forget about it). Operating with
> native lisp arrays is very desired.
>
> GSLL provides good statistics support, but there are some problems
> with contacting to native arrays.

Statistics? Speaking as a research statistician, it sounds like you
want mathematics and probability, not statistics....

but what exactly do you want to do?
 |  Next  |  Last
Pages: 1 2
Prev: Accepting online payment
Next: Question about LLA: how to access a lisp-array