|
|
Re: computers
Gessler, Nicholas (gessler@ANTHRO.SSCNET.UCLA.EDU)
Tue, 7 Mar 1995 20:38:00 PST
Speaking of computers and the National Security Agency, Marc Damashek
under the NSA's auspices, has just published in SCIENCE (10 Feb 95) a most
interesting article on the language-independent fault-tolerant classification
of texts using n-graphs. N-graphs are sequences of n characters which are
tabulated, clustered, and correlated between texts. For me the surprising
result is that by (what he has off-line referred to as "ignorance based
processing") using such a simple algorithm, his software produces quite
meaningful clusters and what might be called abstracts for articles within
the clusters.
Nick Gessler
UCLA - Anthropology
|