Learning Deterministically Recognizable Tree Series
Frank Drewes, Heiko Vogler · 2007
We devise a learning algorithm for deterministically recognizable tree series where the weights are taken from a commutative group. For this, we use an adaptation of the minimal adequate teacher model that was originally introduced by Angluin. The algorithm runs in polynomial time and constructs the unique minimal deterministic bottom-up finite state weighted tree automaton that recognizes the tree series in question.