An ontology learning technique usually starts with formalizing data with a set of features indicative of ontological knowledge, then generates internal representations of candidate models from the formal input of data, and finally applies the best models selected based on certain criteria to produce ontological knowledge.