A Method for Generating History Questions using LOD and Its Evaluation

抄録

The objective of this research is to propose a method to create natural language history questions using current LOD. By combining LOD with a history domain ontology and a history domain dependent question ontology specifying a domain-independent taxonomy, our method can generate content-dependent questions. To be able to support learning, the quality of the questions need to be high. We evaluated the quality of the generated questions by asking a human expert to compare them to questions created by human. The results showed that the system generated question could cover most (84%) of the question supporting basic knowledge acquisition. The results also confirmed that the system generated questions could enhance historical thinking with the same average quality as human generated questions.