Learning of terminal distinguishable languages

We propose new efficient learning algorithms for certain subclasses of regular and even linear languages based on the notion of terminal distinguishability introduced by Radhakrishnan and Nagaraja. The learning model we use is identification in the limit from positive samples as proposed by Gold and further studied by Angluin and many others. All classes we introduce in this paper are proper supersets of the language families TDRL (terminal distinguishable regular) and TDELL (terminal distinguishable even linear) defined by Radhakrishnan and Nagaraja. Extending the classes of efficiently learnable languages is also important from the viewpoint of applications of the algorithms. Some of those extensions are obtained by making use of the concept of control language which is known from formal language theory and has been employed for learning theoretic purposes in particular by Takada. Finally, we are able to exhibit relations to 'merging state' algorithms known for inferring finite automata from positive samples. (orig.)SIGLEAvailable from TIB Hannover: RR 4367(99-23) / FIZ - Fachinformationszzentrum Karlsruhe / TIB - Technische InformationsbibliothekDEGerman