varnamproject-discuss
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Varnamproject-discuss] Improve the learning system


From: Navaneeth K N
Subject: Re: [Varnamproject-discuss] Improve the learning system
Date: Mon, 10 Mar 2014 10:08:58 +0530
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.7; rv:24.0) Gecko/20100101 Thunderbird/24.3.0

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Hello Jaydeep,

On 3/9/14 9:17 PM, Jaydeep Singh wrote:
> Some questions in which your help is required-
> 
> =>I've written a c program which takes hindi words as input and after
> stemming by removing affixes gives the root word. I want to test it
> with varnam. What should i do if i want to print the root along with
> the transliterated word. A function print_token _array() takes token
> which is a pointer to a sturct of type varray and prints it. I'm not
> getting its definition. Also how can I implement my function in
> varnam?

Easy way to test is to edit the `varnam_learn` function and add the
stemming rules there and probably print to the console etc. But since
you have got a C program which does that, you can as well use that.

Proper integration into varnam should go via the scheme file. Which
requires modifying the scheme file, getting the stemming rules compiled
into a VST file, and using it at runtime.

> 
> =>I went through some papers on stemming and found that every language
> would require different sets of stemming rules. Does your idea
> supports this or you want to implement something which can be equally
> used with Indian languages?

Yes this has to be done separately for each languages. But the framework
will be done only once and each language will specify different set of
stemming rules.

> 
> =>I saw the updated ideas page and the idea to create an android ime
> has been luring me.
> Can I submit a proposal for this idea as well?

Please do that. Android IME is high on priority list and people will
certainly love any work on this. This literally enables Indic input on
Andorid devices, which is huge thing IMO. So feel free to explore this
idea and submit a proposal.

> 
> Jaydeep Singh
> (wilderz on IRC)
> 
> On 3/7/14, Jaydeep Singh <address@hidden> wrote:
>> Form my understanding of source code, the vpool here is the initialised
>> resource that we are keeping for use, rather than allocating and destroying
>> again and again as suggested by these function:
>> void*
>> vpool_get(vpool *pool);
>>
>> void
>> vpool_add(vpool *pool, void *item);
>>
>> void
>> vpool_return(vpool *pool, void *item);
>>
>> Program requests an item from the vpool and perform operations on the
>> returned item. When it has finished with an item (or resource), it returns
>> it to the vpool, rather than destroying it. Am I correct in context? Please
>> guide.
>>
> 

- -- 
Cheers,
Navaneeth
-----BEGIN PGP SIGNATURE-----
Version: GnuPG/MacGPG2 v2.0.22 (Darwin)
Comment: GPGTools - https://gpgtools.org

iQEcBAEBCgAGBQJTHUHhAAoJEHFACYSL7h6kCsUH/35OZOyeSWHslYTed75JZcz3
JSazDBCy630qaiYKnH4iDJ21V6DSAsZqVf4n0H1CoM5+1TVlBC56CYfaIceQJev5
8BvbLRIHT86QeDlIDJC9ogsYtUntxc+Y+I8bCbIvfSt+9v7dSGITOfSsCCO6pkDY
IPIqyTq+NgHTlkb7s/bhV+inzAXHDnzhYPRl6C+FLQDFr7JBFilmA5ZA9Uv4X4xg
LfIFteVZKhahFQHpNRIusYyumVWgdmZlauoZDWszv5g9zyDnGnUdBGTvlkIBjkwK
ZoIKGyheOu4zD6yE6qpg1opRwZjdpMSYuvBAn5RlSEQYw6Oof0nAweTvrejeo/A=
=Ogzd
-----END PGP SIGNATURE-----



reply via email to

[Prev in Thread] Current Thread [Next in Thread]