Graphic Showing Associations Between
Abu Sayyaf, Jemaah Islamiyah, and the
National Liberation Front
European News Monitor's Online Explorer Association Program
What follows is a demonstration of data
mining technology. Like much of the tools in the industry, the
quality of your results depends almost entirely on the quality of your
database. For instance, if your database has gleaned unverified
or specifically erroneous data, then what you will see may lead you to
make assumptions that are patently wrong as well. However, if
your data is pretty clean, with multiple sources gathered and also
available in the database, then not only will your output be of high
quality and useful in the life and death world of intelligence, but you
will also be able to see those very same multiple sources as lines from
those sources to the subject. This is critical.
In our example, the European News Monitor provides an online data
fusion and data mining application to anyone who cares to browse their
service. It basically looks up things in their giant database,
and then places the data mining application on top of that,
allowing their website visitor to take a peek at the associations
attributable to the topic at hand. For instance, if you go to the
site and type "Abu Sayyaf" in the search box, you will get a list of
matching entries, and then if you click on the Abu Sayyaf entry you'll
get a list of headlines in the left column and a little graphic in the
right that looks like this. When you click on the graphic you are
automatically routed into the News Explorer Online Association
program. This allows the user to graphically play with the data
to drill down or expand upon items of interest shown. For
instance, the chart shown below demonstrates a rather busy expansion of
the Abu Sayyaf associations in the ENM news database.
Of course not everything in the database implies a nefarious
connection. However the application does allow the user to poke
around and if the data were clean enough, allows the user to make valid
assumptions about the associations found. A news database, of
course, would first need to be gone over carefully removing any
extraneous data...for instance news articles that cover five topics on
the same page might create a false impression that associates people,
places and things from one article into another even when they have
absolutely nothing to do with each other. Or the newpaper or
TV/Radio station itself might be the only link -- it covers a variety
of subjects and as a link, the media is only good for the content, not
the actual linkage itself.
©
Copyright 2007, Michael G. Crawford