Community Ontology

Thursday, February 15, 2007

slide: = Community Ontology = === Puneet Kishor ^1^ and Lewis Friedland ^2^ === ==== University of Wisconsin-Madison ==== ===== 1. Nelson Institute for Environmental Studies ===== ===== 2. School of Journalism and Mass Communication ===== slidecontent: slidehandout:

slide: = What is "Ontology"? = slidecontent:

  • The terms used to describe and represent an area of knowledge
  • Includes … definitions of basic concepts in the domain and the relationships among them
  • Encodes knowledge in a domain and also knowledge that spans domains slidehandout:

slide: = What is "Community"? = slidecontent:

  • An aggregation of organizations, people, locations, and relationships between them
  • Is usually physical, but can also be logical slidehandout:

slide: = "Community Ontology" = slidecontent:

  • The terms used to describe and represent knowledge of…
  • entities and relationships among them in…
  • an aggregation of organizations, people, locations, and relationships between them slidehandout:

slide: = Problem = slidecontent:

  • Lew says, "Puneet, go to the website…
  • identify all the organizations, people, and places, and…
  • deduce the relationships between them
  • Store it all in a repository" slidehandout:

slide: = Solution 1 = slidecontent:

  • Go to the website
  • Visually identify all the organizations, people, and places
  • Visually deduce the relationships between them
  • Store it all in a repository slidehandout:

slide: = Solution 2 = slidecontent:

  • Write a program to crawl the website
  • Download all the webpages and strip out the html tags
  • Open the plain text in an named-entity-tagger (NET) program
  • Manually tag the entities
  • Deduce the relationships between the entities
  • Store it all in a repository slidehandout:

slide: = Solution 3 = slidecontent:

  • Write a program to crawl the website
  • Download all the webpages and strip out the html tags
  • Run the plain text through an automated named-entity-tagger (NET) program
  • Get the results and decide where the precision and recall are satisfactory
  • Deduce the relationships between the entities
  • Store it all in a repository slidehandout:

slide: = Is There A Better Solution? = slidecontent:

  • Yes, let the community tell you its who, what, and where, and their relationships. How?
  • By having them store their data in a format that is readable by other programs, yet
  • Allows them to publish their information as they wish slidehandout:

slide: = Structure Of A Web Page = slidecontent:

<p>Welcome to our community.</p>
<b>Artie MacStrawman</b>
<i>Mayor</i>
<p>Meet your city council.</p>
<ul>
  <li>Councilwoman Elena</li>
  <li>Councilman Joseph</li>
</ul>

slidehandout:

slide: = Structure Of An Intelligent Web Page = slidecontent:

<p>Welcome to our community.</p>
<person>Artie MacStrawman</person>
<title>Mayor</title>
<p>Meet your <agency>city council</agency>.</p>
<ul>
  <li><title>Councilwoman</title> <person>Elena</person></li>
  <li><title>Councilman</title> <person>Joseph</person></li>
</ul>

slidehandout: END