FR3: A Fuzzy Rule Learner for Inducing Reliable Classifiers

Jens Hühn and Eyke Hüllermeier

This paper introduces a fuzzy rulebased classification method called FR3, which is short for Fuzzy Round Robin RIPPER. In the context of polychotomous classification, it uses a fuzzy extension of the well-known RIPPER algorithm as a base learner within a round robin scheme. A key feature of FR3 is its ability to represent different facets of uncertainty involved in a classification decision in a more faithful way, thereby providing the basis for implementing “reliable classifiers” that may, for example, abstain from a decision when not being sure enough.

PDF full paper