The Research Of Automatic Multiple-Choice Questions Generation Based On Ontology

In order to test knowledge level of student in Aircraft Maintenance CBT System, Automatic Item Generation (AIG) is proposed. The research of AIG is not selecting items from item pool, but automatic generate items from teaching books. Multiple-choice questions have advantages of computer operating easily, testing simply and information content richly compares with other types of item, so multiple-choice question is selected as the object of research mainly.Multiple-choice question is composed of the stem and the distractor. A kind of classified method of knowledge points is proposed, and then item pattern is brought forward, it could be used to generate stem. The distractor is generated by the ontology. The ontology can express the terms in the domain and relations among them. The distractors, which have a relationship with the key, could be generated by the ontology. In order to express the level that the distractor affects the key, the distractor level is proposed.A Multiple-Choice Questions Automatic Generation System is developed, in which some primary technology and methods were introduced, for example, regular expression, XML, Jena and so on. Finally, some experiment results of item pattern and the examples of multiple-choice question were given.