| The
keywords are stored in the KeyWord.txt file in four columns. The
first column is a line number. The
second column contains symbols that indicate
that a keyword is synonymous with other words in a group.
The third column contains the keywords, and the fourth has
symbols
under the heading "Property Code". It is mostly reserved for
future
use. A proximity of specific keywords form inputs to a neural network. The number of keywords are also counted in the email to indicate when a keyword occurs freqently. The frequency is compared to a threshold and can provide a script stimulus. The stimulus may start or stop a sequence of scripts. When keywords are used in an algorithm, what is important is not just the keyword, but the keywords plus all it's "synonyms." The word synonym is in quotes because often the related words are not synonyms at all. Equivalence class is a better term. The keyword text file has symbols to indicate equivalence classes and choose an exemplar to represent all the words in it's equivalence class. An example of the keyword text file is shown below. The exemplar has a side arrow ">" pointing to it. Equivalents are coded with an up arrow "^" pointing to the exemplar. The first example in lines 1 - 6 contain equivalence's in a weak sense. The class represents words that are religious, but certainly not synonymous. A second example in lines 8 - 11 gives stronger equivalence's to the extent that they actually are synonyms. The word on line 7 has no equivalences and stands by itself with no symbol necessary. |
| 1 |
> |
god |
x |
| 2 |
^ |
bless |
x |
| 3 |
^ |
christ |
x |
| 4 |
^ |
bible |
x |
| 5 |
^ |
almighty |
x |
| 6 |
^ |
pray |
x |
| 7 |
corn |
x |
|
| 8 |
> |
lawyer |
x |
| 9 |
^ |
solicitor |
x |
| 10 |
^ |
attorney |
x |
| 11 |
^ |
barrister |
x |