Information extraction by means of a generalized k-testable tree automata inference algorithm
Raymond Kosala, Maurice Bruynooghe, Hendrik Blockeel, Jan Van den Bussche · Lirias (KU Leuven) · 2002
We modify a k-testable tree automata inference algorithm so that it performs more generalisation. Evaluation on information extraction tasks for structured documents shows that it performs at least as good as the original.