Prev: defstruct, consistency checking
Next: lisp port
From: S6Carlo on 21 May 2010 10:52
Is there a library available to perform PCA (Principal Comonent
Analysis) in Lisp? I read about Lisp-Stat ( http://www.cs.uiowa.edu/~luke/xls/xlsinfo/
) but I don't know if it has PCA and it looks not so updated.
Thanks,
Carlo
From: allchemist on 21 May 2010 11:11
On 21 май, 18:52, S6Carlo <carlo.savore...(a)gmail.com> wrote:
> Is there a library available to perform PCA (Principal Comonent
> Analysis) in Lisp? I read about Lisp-Stat (http://www.cs.uiowa.edu/~luke/xls/xlsinfo/
> ) but I don't know if it has PCA and it looks not so updated.
> Thanks,
>  Carlo

PCA is a great number of methods. It could be done by SVD, Karhunen–
Loève decomp, NIPALS, some neural networks. What method do you need?

If the words above sounde terrible, simply take SVD (singular value
decomp). There are a lot of information about PCA through SVD. It is
also implemented in several lisp mathematicat libraries. May be GSLL
is a good variant
From: S6Carlo on 21 May 2010 11:18
On May 21, 11:11 am, allchemist <hohlovi...(a)gmail.com> wrote:
> On 21 май, 18:52, S6Carlo <carlo.savore...(a)gmail.com> wrote:
>
> > Is there a library available to perform PCA (Principal Comonent
> > Analysis) in Lisp? I read about Lisp-Stat (http://www.cs.uiowa.edu/~luke/xls/xlsinfo/
> > ) but I don't know if it has PCA and it looks not so updated.
> > Thanks,
> >  Carlo
>
> PCA is a great number of methods. It could be done by SVD, Karhunen–
> Loève decomp, NIPALS, some neural networks. What method do you need?
>
> If the words above sounde terrible, simply take SVD (singular value
> decomp). There are a lot of information about PCA through SVD. It is
> also implemented in several lisp mathematicat libraries. May be GSLL
> is a good variant

Probably we are going to use SVD, but I will do some test in Matlab
before using LISP.
I will look GSLL, but is there a complete implementation of PCA or I
have to write it?
Thanks
From: Tamas K Papp on 21 May 2010 11:42
On Fri, 21 May 2010 08:18:23 -0700, S6Carlo wrote:

> On May 21, 11:11 am, allchemist <hohlovi...(a)gmail.com> wrote:
>> On 21 май, 18:52, S6Carlo <carlo.savore...(a)gmail.com> wrote:
>>
>> > Is there a library available to perform PCA (Principal Comonent
>> > Analysis) in Lisp? I read about Lisp-Stat
>> > (http://www.cs.uiowa.edu/~luke/xls/xlsinfo/ ) but I don't know if it
>> > has PCA and it looks not so updated. Thanks,
>> >  Carlo
>>
>> PCA is a great number of methods. It could be done by SVD, Karhunen–
>> Loève decomp, NIPALS, some neural networks. What method do you need?
>>
>> If the words above sounde terrible, simply take SVD (singular value
>> decomp). There are a lot of information about PCA through SVD. It is
>> also implemented in several lisp mathematicat libraries. May be GSLL is
>> a good variant
>
> Probably we are going to use SVD, but I will do some test in Matlab
> before using LISP.
> I will look GSLL, but is there a complete implementation of PCA or I
> have to write it?

As allchemist said, PCA is not a single method, it is a family of
methods. I don't think that GSL/GSLL includes any of these, but it
should be easy to implement your own. If you don't know how, look it
up in a textbook, or google for SVD PCA.

Tamas
From: allchemist on 21 May 2010 11:44
> Probably we are going to use SVD, but I will do some test in Matlab
> before using LISP.
> I will look GSLL, but is there a complete implementation of PCA or I
> have to write it?
> Thanks

Really, the SVD itself is a complete implementation of PCA, with some
remarks. Simply take US from USV^t and take first N columns.

No, there seems to be no data mining library in common lisp
 |  Next  |  Last
Pages: 1 2
Prev: defstruct, consistency checking
Next: lisp port