VMs: Clustering by context - promising or not?

["PK#01" <pklist01@xxxxxxxxx>]

I found this:

SenseClusters creates clusters of given text instances (an instance could be
a group of one or more sentences that form a context of the target word)
such that instances grouped together in the same cluster are contextually
more similar to each other than to other instances and thus are more likely
to use same meaning of the target word. Each cluster thus represents a
single word meaning that is shared by the instances belonging to that
cluster.

http://senseclusters.sourceforge.net/

I would dearly like to let this loose on the VMS, but I'm afraid that it is
a hell of a job getting it to work. On the other hand it might be a great
way to distiguish gibberish text from natural language.

______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list