Hi Dimitris
Thanks for your feedback. I will post the answers to your questions here.
*1) How do you plan to make the mapping from Wikipedia Categories to
DBpedia classes? Manually or throught the community (like we do with
mappings.dbpedia.org)?*
May be we can use a semi-automatic approach. Once best candidates for
mapping are identified, DBpedia mapping classes could be
suggested automatically. This suggestion could be based on several factors;
some articles in the article cluster
may already have info-boxes and already mapped classes can be
suggested, may be a string comparison could be done with mapping candidates
and DBpedia classes and matched classes could be suggested.
Then a human manually verify the mapping before it is fully incorporated.
Do you this approach is ok?
*2) You mention that you will use the Stanford NER for type suggestions. Is
there a raason why you don't include DBpedia Spotlight? Spotlight returns
directly DBpedia entities.*
I have mentioned Stanford NER because I'm familiar with it and it can be
trained. Honestly, it didn't come to my mind that DBpedia Spotlight could
be used. My mistake, thanks for pointing this out. I will change it.
*3) here are still a few spelling errors but the level is it is generally
improved and I think it is sufficient for the purposes of GSoC.*
I will go through again, word by word and try to correct the spelling
errors.
On Mon, Apr 29, 2013 at 3:43 PM, kasun perera <[email protected]>wrote:
> Hi Marco
>
> I have updated my proposal, corrected my English mistakes. Please have a
> look.
>
> Thanks
>
>
> On Tue, Apr 23, 2013 at 9:41 PM, Marco Fossati <[email protected]>wrote:
>
>> Hi Kasun,
>>
>> Thanks for the proposal.
>> At a first glimpse, I see many errors in your English, so please take
>> care of that. I think we cannot accept applications containing linguistic
>> errors.
>> With respect to the content, below you can find some comments:
>> 1. GSoC project section
>> 1.1. Please extend the 'post-processing of classified entities' part
>> 1.2. In particular, by 'classification ambiguities' you mean wrongly
>> classified entities, don't you? This is a complex task. Can you provide
>> more thoughts on that?
>> 1.3. The Wikipedia category system is already leveraged by the DBpedia
>> extraction framework. A resource is connected to its categories via the
>> http://purl.org/dc/terms/**subject
>> <http://purl.org/dc/terms/subject>property. Hence, resources are already
>> 'clustered' by category.
>> 1.4. However, finding a mapping between Wikipedia categories and
>> DBpedia classes could be interesting. Please develop this idea.
>> 1.5. You don't mention Freebase at all. Freebase is known to provide a
>> well-defined and fine-grained type hierarchy, so I think you should have a
>> look at that project and come back with some thoughts.
>> 2. Open source projects section
>> 2.1. Karsha seems the most related to the GSoC one. Can you detail more
>> on that?
>> 2.2. More specifically, can you expand on the database of FIBO terms
>> and associated documents part?
>> 2.3. Once the most related project is detailed, you could make a
>> synthesis of the others, i.e., shrink them.
>>
>> OK, that's all. Please come back to us once you have updated the
>> application.
>> Cheers,
>>
>>
>> On 4/23/13 4:31 AM, kasun perera wrote:
>>
>>> Hi Marco and Dimitris
>>>
>>> I have submitted my 1st version of the proposal under the project "Type
>>> inference to extend coverage" to the Google site. Can you please look at
>>> it and give me a feedback, so that I can further improve it.
>>>
>>> Thanks
>>>
>>
>> --
>> Marco Fossati
>> http://about.me/marco.fossati
>> Twitter: @hjfocs
>> Skype: hell_j
>>
>
>
>
> --
> Regards
>
> Kasun Perera
>
>
--
Regards
Kasun Perera
------------------------------------------------------------------------------
Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET
Get 100% visibility into your production application - at no cost.
Code-level diagnostics for performance bottlenecks with <2% overhead
Download for free and get started troubleshooting in minutes.
http://p.sf.net/sfu/appdyn_d2d_ap1
_______________________________________________
Dbpedia-gsoc mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbpedia-gsoc
