An Automated Questions Bank Generation and Assessment Framework for Post Graduate Software

Abstract

The evolution in E-Learning technologies establishes the significance of automation in students’assessment process. Particularly, researchers and practitioners are devising newmethodstoautomatically perform students’ assessments effectively during courses. The development of aquestions bank for students’ evaluation processes like quizzes, exams, etc. is highly important. It is usually performed manually by instructors and therefore, the development of effectivequestions for comprehensive assessment is a very time-consuming activity. Consequently, artificial intelligence techniques are employed to generate questions automatically. Althoughthere exist several state-of-the-art studies for the automated generation of questions bank, theseare mostly content-based techniques where rich contents (e.g., paragraphs, book chapters, etc.)are considered for question bank generation. However, in current e-learning or face-to-faceeducational practices, lectures are usually conducted through PowerPoint slides where limitedcontent or contextual information is available. A detailed literature review of 50 researchpaperson question generation indicates the lack of a framework for automated question bank generationfrom PowerPoint slides. Consequently, the existing question bank generation approaches arehard to apply in a real educational environment where slides are the primary mode of lecture. To tackle this issue, An Automated Questions Bank Generation and Assessment Frameworkareproposed in this MS thesis. The proposed framework will employ Artificial Intelligence(AI)techniques like Natural Language Processing (NLP) etc. to enable the automatic generationofquestions bank from PowerPoint slides. The effectiveness of the proposed frameworkwill beestablished through different Postgraduate level software engineering courses where questionsbank will be generated from lecture slides automatically. Our results show the effectivenessofthe proposed algorithm for the text i.e., PowerPoint slides