Create a rule node, link it to the data source node and edit the rule node options.
Click on the SP Paste button to paste the best rule from the last rule discovery experiment into the list box.
Close the dialog and run the stream from the rule node
The best rule extracted from the discovery node should resemble:
IF (Satisfaction 5.97)
AND (DaysConsultancyThisYear 0)
AND (DaysConsultancyLastYear 5)
THEN Response = YES
a = 174, b = 120, c = 119
Accuracy = 68.39, Coverage = 99.17.
The properties a, b and c
tells us that the rule fired for 174 records in the database and
there are 120 records in the YES class. Of the 174 records for which
the rule fired, 119 actually belong to class YES. These properties
give us an accuracy of 68% and coverage of 99%. For more details on
these terms, see the definitions page.
We also notice that the fields that appear in this rule were also amongst the strongest fields listed in the feature selection results.