Towards learning default rules by identifying big-stepped probabilities

Salem Benferhat (1) , Didier Dubois (1) , Sylvain Lagrue (1) , Henri Prade (1)

1 IRIT-ADRIA - Argumentation, Décision, Raisonnement, Incertitude et Apprentissage

This paper deals with the extraction of default rules from a database of examples. The proposed approach is based on a special kind of probability distributions, called "big-stepped probabilities". It has been shown that these distributions provide a semantics for the System P developed by Kraus, Lehmann et Magidor for representing non-monotonic consequence relations. Thus the rules which are learnt are genuine default rules, which could be used (under some conditions) in a nonmonotonic reasoning system, which can be encoded in possibilistic logic.

Keywords: Probability distribution; Databases; Possibility theory; Logic; Encoding

Domains: Artificial Intelligence [cs.AI]

Salem Benferhat, Didier Dubois, Sylvain Lagrue, Henri Prade. Towards learning default rules by identifying big-stepped probabilities. Joint 9th IFSA Congress and 20th NAFIPS International Conference (IFSA 2001), Jul 2001, Vancouver, Canada. pp.1850-1855, ⟨10.1109/NAFIPS.2001.943834⟩. ⟨hal-03299813⟩