9th International Conference on ICT for Sustainable Development

Authors - Dhaval Patel, Krish Bhikadiya, Nikita Bhatt, Ronakkumar Patel, Trusha Patel
Abstract - In education, the quality of question papers is crucial as assessment depend on the performance of students. Manual assessment can be time-consuming, especially with numerous questions. The quality of a question paper can be measured with various factors including difficulty level of questions, relevance to questions and time requirement to attempt the question. This paper focuses on the analysing the quality of question paper based on the bloom’s taxonomy, which is a framework used to assess the quality of a question paper by evaluating the cognitive processes involved, such as remember, understand, apply, analyze, evaluate and create. The work presented here explored various tokenization methods like BERT and deep learning models like LSTM, Bi-LSTM and BERT. The experiments were conducted using a dataset generated by importing questions from university papers across different courses. The generated results conclude that transformer-based BERT model gives the Higher Accuracy than Other Deep Learning models.