Emmanuel Mayer Inductive learning of chronicles , proceedings of the thirteenth biennal european conference on artificial intelligence~(ECAI'98) , 1998 Résumé A usual technique of supervision and diagnosis of complex dynamic systems is the on-line recognition of chronicles, sets of temporally related events. This paper deals with an automatic learning of chronicles based on techniques used in Inductive Logic Programming~(ILP). The principle of the learning method consists in interpreting the induction problem as an equivalent constraint system and using an appropriate constraint manager to solve it. This method is applied on the automatic acquisition of chronicles as part of the supervision of telecommunication networks and power distribution networks. Ce site est maintenu par René Quiniou au moyen du système Weave développé au sein du projet Caravel : http://www-caravel.inria.fr. Dernière modification : 07-10-2011 09:54:25