AI-Powered MCQ Generator Using NLP and LSTM

#ai #nlp #machinelearning

MCQ Generator: Automate Your Quiz Creation Process

Have you ever wished you could automatically generate multiple-choice questions (MCQs) from your lecture notes, articles, or other textual content? Look no further! Introducing MCQ Generator, a Flask-based web application that leverages Natural Language Processing (NLP) and deep learning to dynamically create quiz questions. This project is perfect for educators, students, and content creators who need a quick way to transform text into interactive quizzes.

Overview

MCQ Generator is designed to process input from various sources—whether it's a URL, manual text, or even a file upload (PDF or TXT)—and generate multiple-choice questions with answer options. It even provides a downloadable PDF of the generated questions, making it easy to share or print quizzes for classroom use.

Key Features

MCQ Generation: Automatically generates multiple-choice questions by processing input text.
NLP & Deep Learning: Uses spaCy for NLP processing and a TensorFlow/Keras LSTM model to understand sentence structures.
File & URL Support: Accepts input via URL, manual text entry, or file uploads (PDF/TXT).
PDF Download: Allows users to download the generated MCQs as a PDF.
User-Friendly Interface: Built with Flask and styled using Bootstrap for a clean, responsive design.

Technologies Used

Flask: Backend framework for building the web application.
BeautifulSoup & Requests: For web scraping and processing content from URLs.
PyPDF2: For extracting text from PDF files.
spaCy: For NLP tasks such as tokenization and sentence segmentation.
TensorFlow & Keras: To build and run an LSTM model for understanding sentence structures.
ReportLab: For generating PDFs of the MCQs.
Bootstrap: For a responsive and user-friendly frontend.

Folder Structure

The project is organized as follows:

your_project_folder/
├── app.py
├── requirements.txt
├── README.md
├── static/
│   └── style.css
└── templates/
    ├── index.html
    ├── mcqs.html
    └── result.html

Here’s how it works:

Text Preprocessing:

Uses spaCy to break text into sentences and extract nouns.

LSTM-Based Model:

A neural network is built using TensorFlow/Keras.

The model is trained to understand sentence structures.

MCQ Generation:

Extracts a key noun from the text and replaces it with a blank.

Uses spaCy word vectors to find similar words as distractors.

Shuffles answer choices and formats them as MCQs.

Installation

Clone the repository:

   git clone https://github.com/your-username/nlp-mcq-generator.git
   cd nlp-mcq-generator

Install dependencies:

   pip install -r requirements.txt

Download the English spaCy model:

   python -m spacy download en_core_web_sm

Run the application:

   python app.py

Generating MCQs

Input Options:

Enter a URL to scrape text, paste your own text manually, or upload a PDF/TXT file.

Select Number of Questions:

Choose how many questions you want to generate.

Generate and View:

Click on Generate MCQs to see your questions.
You can also view detailed results with correct answers and download the quiz as a PDF.

Screenshot

Conclusion

MCQ Generator offers a seamless way to convert text into engaging multiple-choice questions, making it an ideal tool for educators and learners alike. With its modern Flask-based architecture and robust NLP processing, creating quizzes has never been easier!