Database Software

Hi,
I'm implementing Email Center, an eBusiness Suite module that manages emails.
There's an email classification functionality that uses Oracle Text to classify email. I believe this works by using Oracle text to assign scores to an email. I guess it compares the email content (text) with a set of keywords in the Email Center module.
My problem is that I can't figure out what's the logic used to classify the emails, i.e., to calculate the score of each email that's received. I can specify the keyword I want to use for the classification process. I've been performing tests for two days know and, for the sake of me, I can'd understand why emails are classified the way they do.
There's a table that contains the following text and a score:

" about (DOCTYPE)*2.8,about (HTML)*5.5,about (PUBLIC)*2.8,about (W3C)*2.8,about (DTD)*2.8,about (3.2)*2.8,about (EN)*2.8,about (TESTE)*2.8,about (CLASSIFICAÇÃO)*2.8,about (76)*2.8, about (EMPRESA)*5.8,about (STAPLES)*5.8,about (ORACLE)*5.8"

Does this gibberish mean anything to you?
How can I test the scoring process to understand its logic?
How can I fine-tune the scoring process to give the results I want? I couldn’t find an "implementation guide" for Oracle Text.

Any hint you may give me is helpful, since I don’t know a thing about Oracle Text.
Thanks a lot
------------------------------------
João Vargem