Noise detection and elimination applied to noisehandling in a

Compression measures used in inductive learners, such as measures based on the MDL (Minimum Description Length) principle, provide a theoretically justiied basis for grading candidate hypotheses. Compression-based induction is appropriate also for handling of noisy data. This paper shows that a simple compression measure can be used to detect noisy examples. A technique is proposed in which noisy examples are detected and eliminated from the training set, and a hypothesis is then built from the set of remaining examples. The separation of noise detection and hypothesis formation has the advantage that noisy examples do not innuence hypothesis construction as opposed to most standard approaches to noise handling in which the learner typically tries to avoid overrtting the noisy example set. Experimental results in a KRK (king-rook-king) chess endgame domain show the potential of this novel approach to noise handling.