Overview

This project utilises NLP techniques and large language models (LLMs) to analyse the anime series. The goal is to produce a comprehensive dataset, extract themes, classify text, and generate a character network. The end result will be an intuitive web-based GUI created with Gradio.

Project Features

1. Web Scraping (crawler)

• Utilizes Scrapy to extract structured information from Narutopedia.
• Builds a dataset containing essential details about Naruto characters, themes, and dialogues.

2. Character Network (character_network)

• Uses spaCy's Named Entity Recognition (NER) to identify character mentions in dialogues and descriptions.
• Constructs an interactive character network using NetworkX and PyViz, allowing visualization of relationships between characters.

3. Text Classifier (text_classifier)

• Trains a model to classify dialogues into predefined categories.
• Uses deep learning and transformer-based models to achieve high accuracy.

4. Theme Extraction (theme_classifier)

• Implements Zero-shot classification with Hugging Face's transformers to identify key themes from dialogue and descriptions.
• Helps in understanding the overarching topics present in different episodes or story arcs.

Technologies Used

• Python (Primary programming language)
• Scrapy (Web scraping)
• spaCy (Named Entity Recognition)
• NetworkX & PyViz (Graph visualization)
• Hugging Face Transformers (Zero-shot classification, LLMs)
• TensorFlow/PyTorch (Text classification)

Future Enhancements

• Develop a character chatbot using LLMs for interactive conversations.
• Expand web scraping to include more anime series.
• Enhance visualization features for character relationships.