We introduce a rule selection algorithm called ROCCER, which operates by selecting classification rules from a larger set of rules ? for instance found by Apriori ? using ROC analysis. Experimental comparison with rule induction algorithms shows that ROCCER tends to produce considerably smaller rule sets with compatible Area Under the ROC Curve (AUC) values. The individual rules that compose the rule set also have higher support and stronger association indexes.
|Translated title of the contribution||ROCCER: an Algorithm for Rule Learning Based on ROC Analysis|
|Title of host publication||Unknown|
|Pages||823 - 828|
|Number of pages||5|
|Publication status||Published - Aug 2005|