• View
• Edit
• History
• Print

3857

 I want to make sure ItWorks 

General Navigation

• Help Pages | FrontPage | StartingPoints | http:wikiNames -- all pages on this wiki | Changes | RecentChanges | http:quickChanges?days=1 | http://c2.com/cgi/RecentChanges?days=.25

Searches

• Page Search: [Search]
• Word Search: [Fullsearch]
• LocalWiki
  • http://localhost

MyWebSite

• Home
  • http://drn.wiki/default.htm
  • http://drn.wiki/WikiTest/RelevantChanges.htm

Projects

• InfoSpace Development and Construction
  • http://drn.wiki/WikiTest/Project_InfoSpace.htm

Geography

• Map
  • http://www.virtualmex.com/newmap.htm

FastFacts:

• When you search for a page or pages on this wiki, the number of pages reported on the last line is the current number of pages.
  • This was pointed out to me by ElizabethWiethoff. I did a test to confirm this fact by creating a page, checking the count, then deleting the page and checking the count again. The report I got accounted for the changes.
• For a links to blogs of SoftwareThinkers (including our own WardCunningham), try:
  • http://www.artima.com/weblogs/index.jsp
    • On Weblogs -- http://www.artima.com/weblogs/viewpost.jsp?thread=152735

Considerations today: 20061003

A feature to be added to NysInfo: SemanticRelatedness

"Is needle more related to thread than it is to pie? Humans agree on the relatedness of most word pairs. Miller and Charles (1991) suggest that relatedness is based on the overlap of contextual representation of words."

• http://www.d.umn.edu/~tpederse/Pubs/patwardhan.ppt

One of the strengths of the HyperlinkedDocument is the use of WordPairs or WordGroups as page names representing the initial content of the page.

Considered also as valuable are groups of documents which have been collected using such mechanisms of relating pages as that of Categories, topics and RoadMaps. All three of these share in the aspect of having SemanticRelatedness. While humans have the inherent means of making associations based on a sense of recognition of relationships, an automated agent can only approach this ability by considerable programming.

In constructing such an agent, over 200 different techniques and schemes might be employed to simulate this human sense. Converting the techniques and schemes into working code, is not simple process, but given sufficient resources and time, it may be possible to simulate a limited process to relate documents via a relatedness due to the semantics and collections of words within the considered documents.

CategoryHomePage