This project involves the analysis of dialogues from the Game of Thrones (GoT) series using NLP techniques. The main objective is to explore how dialogues relate to characters in the series and derive insights from these relationships. Additionally, the project investigates the feasibility of using neural networks to generate dialogue scripts.
Ensure you have the following Python libraries installed:
- numpy
- pandas
- matplotlib
- scikit-learn
- seaborn
- fastai
- networkx
- nltk
- gensim
Raw dialogues were obtained from Genius.com. The dataset consists of CSV files containing character dialogues, seasons, and chapters. For detailed data cleaning and preprocessing steps, refer to the Final_Pipeline notebook.
- Final_Pipeline.ipynb: Comprehensive data preparation, analysis, modeling, and scoring process.
- GoT dialogue generator.ipynb: Use a pre-trained model to generate random GoT dialogues.
- GoT: Raw chapter scripts obtained from Genius.com.
- CSV: Cleaned CSV files containing character dialogues.
Character interactions were visualized using a network graph generated from the dialogue data.
Dialogue generation using GPT-2 language model is demonstrated in the GoT dialogue generator.ipynb notebook.
For model training, refer to the instructions in the runmeforallwork.ipynb notebook.
- Md. Sajid Ahmed - 1610364042
- Zahin Akram - 1610618042
- Arifuzzaman Arman - 1610551042
- Md. Rakib Imtiaz - 1610294042