...texts[APBL91].
In the papers we found, the system was not finished completely, and could not deal with large volumes of text.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...contents.
This way of learning could be said to correspond to simple SELECT sentences with projections in SQL (see for instance [DD93] or [EN94]) used in most relational database systems.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...condition
The attributes are divided into condition attributes and decision attributes. Condition attributes are the attributes the classification is based on, while the decision attributes are what distinguish the classes, i.e. the classification.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...4#4
This means: If Color is red and Size is big, then the type is Pelican.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...stage.
The discernibility matrix given is not possible in practice, but serves well for illustrative purposes it.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...sets.
When reading this subsection, bear in mind that the article [Szl94] which this is based on is rather outdated (winter 1994), and that the situation is likely to have changed.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...format
The format of the table file is a text file with tabs as column separators.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...RSES
To generate the files, type rses& and choose Editor and Preprocessor. Then open the file, File-Import and save it again, File-Save As. Save both the decision table and the dictionary file.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

Helge Grenager Solheim
Sat May 4 03:30:02 MET DST 1996