Digital Libraries Projects - ET-Space
ET-Space - Visualizing Entertainment
Introduction
In 1995, a testbed of 110,000 Internet homepages from the entertainment section of Yahoo! was gathered by an Internet Spider. An automatic indexing algorithm was applied to the homepages and a concept space (ET-Space) and multi-layered Kohonen SOM (ET-Map) were created.
ET-Space:
The Concept Space interface allowed
users to search for documents directly
with keywords or access the concept
space for search term suggestion.
The concept space was an automatically
generated thesaurus based on word
co-occurrence both within individual
documents and within the collection
as a whole. The system could suggest
terms for searching as they appear
in the documents enhancing information
recall. The top 40 terms were displayed
in order based on a weighting/ranking
algorithm. Each concept space term
had an identifier that links it to
the source search term. The interface
allowed the user to enter search terms
as a single word or a phrase and multiple
search terms could be entered. Clicking
on the box adjacent to a concept space
term added it to your previous
search. The user could then elect to see
more concept space terms or search
for documents.
ET-Map:
A 2-D multi-layered SOM was generated
from the entertainment collection.
Text labels and colors are used to
demarcate regions in the SOM, color
has no specific meaning. SOMs will
cluster similar categories near each
other on the map. Clicking on a map
region will take you down a layer
in a multi-layered map, or show you
the documents associated with that
category if the number of documents
is less than 200.




