Wikipedia Q/A Chatbot

This project is a Wiki Q/A Chatbot that combines interactive chit-chat with topic-specific information retrieval. It is built using a dataset of over 50,000 Wikipedia documents and provides concise, relevant answers across topics like health, technology, politics, and more.

Features

• Chit-chat mode powered by BlenderBot for natural conversations.
• Topic classification using zero-shot learning to ensure accurate responses.
• Document retrieval and summarization with TF-IDF for efficient Q/A.
• Exception handling to maintain smooth user interactions.
• Web-based chat interface with topic selection and conversation control.

Methodology

• Data Collection: Wikipedia scraping for 10 major topics (e.g., health, environment, technology).
• Processing: Preprocessing and cleaning of scraped documents.
• Chit-chat Component: Integrated BlenderBot for engaging, dynamic conversations.
• Q/A Retrieval: TF-IDF vectorization and summarization for precise answers.
• Topic Analysis: Zero-shot classification using NLP models.

Tech Stack

• Languages: Python
• Libraries: BlenderBot, TF-IDF, BERT-based zero-shot models
• Frameworks: Flask (web interface)
• Data Source: Wikipedia (via wikipedia library)

Conclusion

The chatbot integrates conversational AI with information retrieval, providing both casual and topic-specific interactions. It is designed for accuracy, efficiency, and a smooth user experience.