>
> I surely wish there were some sort of software which would read in a
> file and provide an expert system's "opinion" on what categories the
> file would emcompass...
Don't know if the following will help (just doing dogpile searches
for capital expenditure next year, and since I was already in the
search engine...) but I found a place with text anaylsis tools.
http://etext.lib.virginia.edu/textual.html
"DiscAn
DiscAn bills itself as a content analysis program; among its many
other features, it can compare two corpora of works and find points of
similarity and discontinuity, complete with frequencies and
percentages. It also provides a "discourse analysis component" --
which has been used to predict the acceptability of possible metaphors
in a given culture, and analyse the relative dynamics of pornography
for the Canadian Ministry of Justice. "
"WordCruncher
Recently resurrected, WordCruncher is a text reader/analysis program
that does not read ASCII files directly, but rather generates an index
from that file to read. Recognizing this subtle distinction makes it
easier to understand WordCruncher's "15,000 word limit;" it is not
simply words that WordCruncher counts, but unique words. The entire
corpus of an author, for instance, can be read under WordCruncher,
provided the author used less than 15,000 different words in all
his/her works.
Once texts are indexed, WordCruncher enables the user to search large
texts easily, as well as generate keyword-in-context (KWIC)
concordances, word frequency lists, and what WordCruncher calls
"book-style" indexes. http://www.wordcruncher.com "
-Coyt D. Watters
"The Internet, billions of electrons with nothing better to do."