It's possible to classify examples using rules previously learned. Yails gives the user two possibilities. Classify a file of examples or give it an examples manually. An example interaction showing the commands for these two cases follows :
YAILS - classify example
Existing attributes :
Attr no.1 => color,symbol Attr no.2 =>
power,numeric
Attr no.3 => cilind,numeric Attr no.4 =>
make,symbol
Attr no.5 => max_speed,numeric Attr no.6 =>
accel,numeric
Attr no.7 => country,symbol
Introduce the example in the format : ValueAttr1 ValueAttr2 ...
You can use values "dontcare" and "unknown"
EXAMPLE :: black 88 1490 alfaromeo 175 11 italy
Predicted Class : family
YAILS - classify file toy_stat
GLOBAL RESULTS
Correct Answers -> 16
Incorrect Answers -> 0
Unknown Answers -> 0
Accuracy (%) -> 1
Number of used rules -> 6
Average number of selectors/rule -> 1.333
1 Rules in the Background Theory
On the first case we have the command for manual introduction of examples (classify example). Yails gives as result the predicted class of the example. If the user chooses to classify a file with examples then Yails calssifies them comparing its answers with the ones present in the file, presenting the global results to the user at the end of the classifcation task.
Internally Yails obtains the set of rules wich could be used to classify each example. Then as said in section 3.3 the "best rule" is choseen. This version of the interface only uses this best answer. Presently a new interface based on X-windows toolkit available for Yap-prolog is being developed. This new version presents the user with the best 10 alternatives to the outputed answer togheter with confidence (the OV values) on these alternatives.