Question generation system

ABSTARCT :

The prospect of applying the semantic relationships to the question generation system can revolutionize the learning experience. The task of generating questions from the existing information is a tedious task. In this paper, Question generation system based on semantic relationships (Q-Genesis) is proposed to generate more relevant knowledge level questions automatically. It will be useful for the trainer to assess the knowledge level of the learners. This paper also provides the importance of the semantic relationships when generating the questions from the ontology.

EXISTING SYSTEM :

? Existing research towards question generation systems is widely based on template and syntax. Recently, the researchers focused towards semantics based question generation system. ? The semantic relationship that exists between concepts provides common understanding of the knowledge. Ontologies provide open world semantics and make domain assumptions explicitly. ? Semantic relationships that exist in the ontology provide more semantics for an application. It is widely used to accommodate real world knowledge. ? In this work, the sematic relationships that exist between the concepts are exploited to generate more number of knowledge level questions.

DISADVANTAGE :

? We found that manually created questions exhibit the highest DP and there is no statistically significant difference between our system and the state-of-the-art system, implying that questions generated by our system are as good as, if not better than, questions generated by the state-of-theart system. ? A number of issues need to be addressed in future research. ? The feasibility of automatically or semi-automatically choosing pre-questions needs to be explored. ? A large-scale experiment investigating the productivity of generating questions (time taken to post-edit questions vs. time taken to generate questions from scratch) is planned.

PROPOSED SYSTEM :

• In this paper, Question generation system based on semantic relationships (Q-Genesis) is proposed to generate more relevant knowledge level questions automatically. • Self-questioning strategy was proposed to generate the questions from narrative sentences. • They have used OpenLearn data resource for question generation. proposed a framework to generate comprehension questions automatically. • The logical form can be used to generate WH questions. proposed an approach to automatically generate multiple choice questions.

ADVANTAGE :

? We analyse the effect such pre-questions have on test-takers’ performance on a comprehension test about a scientific video documentary. ? QG systems help promote student learning by providing learning content and forms of assessment which allow for convenient and fast evaluation of student performance. ? However, when comparing the performance of students who received text-based pre-questions with that of those who received pre-questions with screenshots, we found no statistically significant difference (p=0.1537). ? One measure is item discriminating power (DP) (Gronlund, 1982). DP describes the relationship between student performance on a particular item and their total exam score.