[Corpora-List] Fwd: The standard size of splitting the dataset

[Michele Filannino]

I share my answer.

---------- Forwarded message ----------
From: Michele Filannino <michele.filannino at cs.manchester.ac.uk>
Date: Thu, Jun 27, 2013 at 2:31 PM
Subject: Re: [Corpora-List] The standard size of splitting the dataset
To: Jack Alan <j.o.alan2012 at gmail.com>


Hi Jack,

the question is exhaustively addressed in the attached paper.

Bye,
michele.



On Thu, Jun 27, 2013 at 1:58 PM, Jack Alan <j.o.alan2012 at gmail.com> wrote:

> Hi all,
>
> Has anyone came across the standard size of splitting the dataset
> into (training, development and test) in supervised learning? I mean what
> is the typical percentage size for each subset especially for sequence
> labelling tasks, e.g. POS and NER?
>
> I wonder if it is something like 60% training, 20% development and 20%
> test?
>
> Many thanks
>
> _______________________________________________
> UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
> Corpora mailing list
> Corpora at uib.no
> http://mailman.uib.no/listinfo/corpora
>
>


-- 
Michele Filannino

CDT PhD student in Computer Science
Room IT301 - IT Building
The University of Manchester
filannim at cs.manchester.ac.uk



-- 
Michele Filannino

CDT PhD student in Computer Science
Room IT301 - IT Building
The University of Manchester
filannim at cs.manchester.ac.uk
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20130628/a418c681/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 10.1.1.33.1337.pdf
Type: application/pdf
Size: 202283 bytes
Desc: not available
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20130628/a418c681/attachment-0001.pdf>
-------------- next part --------------
_______________________________________________
UNSUBSCRIBE from this page: http://mailman.uib.no/options/corpora
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora