RE: [newsml-2] RE: Event significance

Dear Misha,

So, Katrina-the-event is associated with Meteorological-disaster-the-theme and
Flood-the-theme. It is also associated with New Orleans-the-city, which is
naturally associated with Louisiana-the-state. Finally it is also associated
with oil-industry-the-business-sector.

2 questions: 
a) what is the scope of the associations your journalists think about? The
global scope of your knowledge base? Or the local scope a given news item?

b) In the latter case, apart from academic interest, do you think that
journalists will spend time creating these associations on the fly? And what is
the interest of this supplemental information for a *consumer* of news items?

In the first case, I believe that this kind of relationship should be part of
the semantic network a news provider may create in its knowledge base. The
*concept links* we have drafted as part of the topic Item model should be used
for this purpose. These associations are interesting for semantic navigation
purposes on a news provider's platform. As topic Items may be distributed to
news customers, these associations may be processed, not only by their creators,
but in fact by all newsml2 customers along the news industry chain. 

 <topicItem>
	xxxxxxx
      <title xml:lang="en">Katrina disaster</title>
	xxxxxx
      <conceptId code="evt:Katrina />
      <link rel="relatedTheme" code="nc:03007000">
		<title xml:lang="en">Meteorological disaster</title>
	</link>
      <link rel="relatedTheme" code="nc:03005000">
		<title xml:lang="en">Flood</title>
	</link>
      <link rel="relatedLocation" code= iso3166-2:US-LA ">
		<title xml:lang="en">Louisiana</title>
	</link>
	xxxxxxx
<topicItem>

(note: namespaces skipped in the sample)

These concept associations could then be given in news Items as hints, like in:
 
    <subject code="evt:Katrina>
      <title xml:lang="en">Katrina disaster</title>
      <link rel="relatedTheme" code="nc:03007000">
		<title xml:lang="en">Meteorological disaster</title>
	</link>
      <link rel="relatedTheme" code="nc:03005000">
		<title xml:lang="en">Flood</title>
	</link>
      <link rel="relatedLocation" code= iso3166-2:US-LA ">
		<title xml:lang="en">Louisiana</title>
	</link>
    </subject>

What you may miss here is an indication of the strength of the association (what
you call the significance). In neural (neuronal) networks, such a value exists,
as a weight: maybe we should add this information to the concept-link property.

      <link rel="relatedLocation" weight="80" code= iso3166-2:US-LA ">
		<title xml:lang="en">Louisiana</title>
	</link>

Would this approach fit you needs?

Laurent


-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

This e-mail, and any file transmitted with it, is confidential and  intended solely for the use of the individual or entity to whom it is addressed. If you have received this email in error, please  contact the sender and delete the email from your system. If you are  not the named addressee you should not disseminate, distribute or copy  this email.

For more information on Agence France-Presse, please visit our web site at http://www.afp.com

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Received on Monday, 13 March 2006 16:54:06 UTC