The current perception is that a couple of paragraphs mentioning the entity
seem to be good enough.

For a more rigorous answer, you may have to wait for the experiments that
we are running for a paper that we're working on, where we actually test
several hypothesis.

Cheers,
Pablo

On Fri, Mar 8, 2013 at 9:11 AM, Chux <[email protected]> wrote:

> Hi Pablo,
>
> I guess I'm looking for the optimal amount of supporting text for
> spotlight to annotate correctly.
>
> Say in automated environments where the amount of text is variable, you'd
> want some threshold for the length of the context to not overload the
> spotlight web service.
>
> Thanks,
> Chux
>
> On Thu, Mar 7, 2013 at 8:55 PM, Pablo N. Mendes <[email protected]>wrote:
>
>> Hi Chux.
>> Why not the entire paragraph?
>>
>> Cheers,
>> Pablo
>>
>> On Fri, Mar 8, 2013 at 1:29 AM, Chux <[email protected]> wrote:
>>
>>> Hi Guys,
>>>
>>> I asked before about "manual spotting" and as follow up to that, I was
>>> wondering what would be the most effective placement of a "Spot" to be
>>> disambiguated within a context for Spotlight to make the most accurate
>>> disambiguation, since spotlight is looking at the context of a word before
>>> it is annotated. Is it in the beginning of a context, middle or end?
>>>
>>> For example on the same case:
>>>
>>> *(1) Windows. For the same $1,299 you could spend on this 21.5-inch
>>> Apple iMac, you can buy a 23-inch or even a 27-inch (2) Windows-based
>>> all-in-one, and with more robust features.*
>>> *Yes, for better or for worse, those other PCs have (3) Windows 8. They
>>> also have touch screens and optical drives, features that Apple has deemed
>>> inessential for a desktop. You might agree with Apple. You might also
>>> accept Apple's hallmark design excellence and display quality as worthy
>>> substitutions. You might even need or want a particular function of OS X
>>> that you can't get in a (4) Windows PC.*
>>>
>>> So for example I chose to manually annotate *(3)*, will it give me a
>>> higher chance of having a correct disambiguation of  *"Windows"*compared to 
>>> picking
>>> *(1)*,* (2)*, or *(4)*?
>>>
>>> Essentially what I'm looking to get to is how much context do I need to
>>> feed spotlight (in my case since I wanted a manual way to spot words) for
>>> it to be accurate.
>>>
>>> Thanks,
>>> Chux
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester
>>> Wave(TM): Endpoint Security, Q1 2013 and "remains a good choice" in the
>>> endpoint security space. For insight on selecting the right partner to
>>> tackle endpoint security challenges, access the full report.
>>> http://p.sf.net/sfu/symantec-dev2dev
>>> _______________________________________________
>>> Dbp-spotlight-users mailing list
>>> [email protected]
>>> https://lists.sourceforge.net/lists/listinfo/dbp-spotlight-users
>>>
>>>
>>
>>
>> --
>>
>> Pablo N. Mendes
>> http://pablomendes.com
>>
>
>


-- 

Pablo N. Mendes
http://pablomendes.com
------------------------------------------------------------------------------
Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester  
Wave(TM): Endpoint Security, Q1 2013 and "remains a good choice" in the  
endpoint security space. For insight on selecting the right partner to 
tackle endpoint security challenges, access the full report. 
http://p.sf.net/sfu/symantec-dev2dev
_______________________________________________
Dbp-spotlight-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dbp-spotlight-users

Reply via email to