Machine learning support

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

Machine learning support

sharanbr
Hi,

I would like to know if there are efforts to add Machine learning features
or packages in Octave.
I am on my way to looking at application of ML for my  area of work. The
people I talk to all seem
to suggest that R/Python are the best languages for ML.

I want forum members take on this.




--
Sent from: http://octave.1599824.n4.nabble.com/Octave-General-f1599825.html

_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Machine learning support

Juan Pablo Carbajal-2
Hi,

I work on ML mainly using Python (scikitlearn, mltk, theano, etc) and
Octave (gpml, stk, optim, etc). I stay away from R when possible. I
also use C/C++ (Shogun, MLPack, etc...) which many times can provide
an oct file interface.

ML as a topic is too wide. You rather focus on one aspect of ML and
write your package, examples are nnet, fuzzy, gpml.
If your plan is to do ML on text, I would say that Octave is not the
best choice, in this cas eI woudl stick to Python.

Take a look at Julia language as well, there is plenty of packages
already for that one.

Cheers

_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Machine learning support

sharanbr
THanks. I have another question.
As I am new to ML, I am looking for some inputs about the right method to
use for a problem statement.
However, I have had little help from redditt forum on this.

Do you know a forum where people are more active  and willing to help?





--
Sent from: http://octave.1599824.n4.nabble.com/Octave-General-f1599825.html

_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Machine learning support

stimits
If you are using NVIDIA CUDA-based machine learning see:
https://devtalk.nvidia.com/default/board/53/accelerated-computing/
 
----- Original Message -----
From: sharanbr <[hidden email]>
To: [hidden email]
Sent: Fri, 29 Dec 2017 03:45:19 -0000 (UTC)
Subject: Re: Machine learning support
THanks. I have another question.
As I am new to ML, I am looking for some inputs about the right method to
use for a problem statement.
However, I have had little help from redditt forum on this.
Do you know a forum where people are more active and willing to help?
--
Sent from: http://octave.1599824.n4.nabble.com/Octave-General-f1599825.html
_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave

_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Machine learning support

Carlo de Falco-2
In reply to this post by Juan Pablo Carbajal-2


> On 28 Dec 2017, at 16:08, Juan Pablo Carbajal <[hidden email]> wrote:
>
> Hi,
>
> I work on ML mainly using Python (scikitlearn, mltk, theano, etc) and
> Octave (gpml, stk, optim, etc). I stay away from R when possible. I
> also use C/C++ (Shogun, MLPack, etc...) which many times can provide
> an oct file interface.
>
> ML as a topic is too wide. You rather focus on one aspect of ML and
> write your package, examples are nnet, fuzzy, gpml.
> If your plan is to do ML on text, I would say that Octave is not the
> best choice, in this cas eI woudl stick to Python.
>
> Take a look at Julia language as well, there is plenty of packages
> already for that one.
>
> Cheers

As far as NNET is concerned, there has been a recent GSOC project intended
to revive the NNET package for Octave last year.

The student and mentor of this project are in CC.

c.
_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Machine learning support

Juan Pablo Carbajal-2
In reply to this post by sharanbr
> I am looking for some inputs about the right method to
> use for a problem statement.

The question is too general and vague. Be specific and answer will come.

_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Machine learning support

Alois Schloegl-7
In reply to this post by Juan Pablo Carbajal-2
On 2017-12-28 16:08, Juan Pablo Carbajal wrote:

> Hi,
>
> I work on ML mainly using Python (scikitlearn, mltk, theano, etc) and
> Octave (gpml, stk, optim, etc). I stay away from R when possible. I
> also use C/C++ (Shogun, MLPack, etc...) which many times can provide
> an oct file interface.
>
> ML as a topic is too wide. You rather focus on one aspect of ML and
> write your package, examples are nnet, fuzzy, gpml.
> If your plan is to do ML on text, I would say that Octave is not the
> best choice, in this cas eI woudl stick to Python.
>
> Take a look at Julia language as well, there is plenty of packages
> already for that one.
>
> Cheers
>
> _______________________________________________
> Help-octave mailing list
> [hidden email]
> https://lists.gnu.org/mailman/listinfo/help-octave
>


Hi Juan,


I recommend taking a look at the "NaN-toolbox - A statistics and machine
learning toolbox ..." [1].

It contains a number of machine learning methods. For more details see
the help functions of train_sc, test_sc, classify, and xval:

  https://octave.sourceforge.io/nan/function/train_sc.html
  https://octave.sourceforge.io/nan/function/test_sc.html
  https://octave.sourceforge.io/nan/function/classify.html
  https://octave.sourceforge.io/nan/function/xval.html


Cheers,
  Alois



[1] The NaN-toolbox: A statistics and machine learning toolbox for
Octave and MatlabĀ® for data with and w/o MISSING VALUES encoded as NaN's.

https://pub.ist.ac.at/~schloegl/matlab/NaN/
https://octave.sourceforge.io/nan/index.html
https://mloss.org/software/view/206/

_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Machine learning support

Juan Pablo Carbajal-2
> I recommend taking a look at the "NaN-toolbox - A statistics and machine
> learning toolbox ..." [1].

I use the toolbox (for nan working functions like mean, etc) but a
classifier hardly makes it a ML toolbox. As said, ML is too wide (and
I would say an over claim for the nan toolbox).
Thanks

_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Machine learning support

sharanbr
In reply to this post by Juan Pablo Carbajal-2
Hi Juan,The following is my problem statement (in bold):*The system i am
testing basically generate events. Sequence of these are of interest to
me.One can manually look at these event sequence and recognize them to be
useful. But manual process is very cumbersome and also there could be
millions of events within which one has to look for interesting events.The
interesting event sequence are known a-priori. The spacing between these
events can vary though.*I can provide more details if needed.> I am looking
for some inputs about the right method to> use for a problem statement.The
question is too general and vague. Be specific and answer will come.



--
Sent from: http://octave.1599824.n4.nabble.com/Octave-General-f1599825.html

_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Machine learning support

Juan Pablo Carbajal-2
You are looking for pattern matching
https://en.wikipedia.org/wiki/Pattern_matching

_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave
Reply | Threaded
Open this post in threaded view
|

Re: Machine learning support

Juan Pablo Carbajal-2
On Mon, Jan 1, 2018 at 1:06 AM, Juan Pablo Carbajal
<[hidden email]> wrote:
> You are looking for pattern matching
> https://en.wikipedia.org/wiki/Pattern_matching

or if your patterns aren't as exact as you describe them, then
https://en.wikipedia.org/wiki/Pattern_recognition

_______________________________________________
Help-octave mailing list
[hidden email]
https://lists.gnu.org/mailman/listinfo/help-octave