Smart MCQ Generation using KeyBERT and BERT Based Models

This project proposes an Automated Multiple-Choice Question (MCQ) Generator that generates questions by analyzing text content using natural language processing (NLP) techniques. The system preprocesses text, extracts keywords, and associates them with appropriate sentences using Transformers, KeyBERT, and NLTK. Significant keywords are extracted by KeyBERT's BERT-based sentence embeddings, enabling direct sentence selection. Relevant sentences are chosen using KeyBERT's BERT-based sentence embeddings, which extract important keywords. To guarantee semantic accuracy, the system predicts word senses using a BERT-based MLM. MCQ generation is facilitated by blank substitution in sentences, which provides both correct answer and incorrect answers. User responses are saved for scoring and feedback. This useful application enables educators, makers, and students to quickly and efficiently create comprehension examinations. The system receives text files from users, which ensures flexibility and adaptation to a variety of educational circumstances and content domains.