Automatic question generation and answer assessment for subjective examination

Automatic generation of questions and evaluating their answers is a highly challenging task in natural language processing and educational technology. This work focuses on generating subjective questions and also an evaluation system is suggested for assessing the answers. For generating the questionnaires, key-phrases are extracted from the course curriculum (syllabus). Next, based on the key-phrases, different types of subjective questions are generated. Finally, the evaluation of student’s responses is achieved using a multi-criteria-decision-making approach. It uses a set of model answers taken from different textbooks and subject experts to evaluate the answers. Multiple measures are used to assess the answers by comparing them with this model set. The results of the profound system reveal that the automated appraisal process can reduce the manual effort of the human.