AI Question-Answering Chatbot

An intelligent chatbot system powered by BERT NLP model with text interaction capabilities.

✨ Key Features

• BERT-based Q&A: Utilizes deepset/bert-base-cased-squad2 model
• Dual Interface: Web-based text chat interaction
• Admin Dashboard: For content management and customization
• High Accuracy: Achieved 80.11% F1 score on validation data
• Easy Integration: Flask-based backend with responsive Bootstrap frontend

🛠 Technical Specifications

Model Performance

Metric	Score
F1 Score	80.11%
Precision	77.50%
Recall	88.57%

Technology Stack

Backend:

• Transformers (BERT Q&A)
• PyTorch
• Flask

Frontend:

• Bootstrap 5
• JavaScript

Installation & Setup

1. Clone the repository

   git clone https://github.com/rahmaeBrahi/Chatbot-NLP.git cd Chatbot-NLP

2. Install dependencies pip install -r requirements.txt

3. Initialize the model python main.py --init-model

4. Run the backend python backend.py

5. Configuration Edit config.json to customize settings

6. Advanced Options Model Switching: Replace model name in main.py: MODEL_NAME = "deepset/bert-base-cased-squad2" Change to roberta-base or other HuggingFace models if needed.

Database Integration: Pre-configured for PostgreSQL/MongoDB

7.👩‍💻 Development Team

Rahma Ebrahim

Poussy Ayman

jehad mahmoud