COMP90042-Rumour-Detection-on-Twitter

This repository contains the source code for the Rumour Detection and Analysis on Twitter Project that was part of the COMP90042 Natural Language Processing course at the University of Melbourne.

Project structure

• data/ -- Raw datasets published by COMP90042 competition organizers
• doc/ -- Documentation and project report (LaTeX source)
• src/ -- Source code for task 01 (rumour identification) and task 02 (rumour analysis)
  • 01_rumour_detection_bertweet.ipynb -- Notebook using pre-trained BERTweet model
  • 01_rumour_detection_multimodal_bert.ipynb -- Notebook with implementation of Multimodal Toolkit architecture
  • 01_rumour_detection_tf_with_huggingface_model_hub.ipynb -- Notebook using pre-trained BERT models from Hugging Face Model Hub
  • 01_rumour_detection_tf_with_tf_hub.ipynb -- Notebook using pre-trained BERT models from Tensorflow Hub
  • 02_rumour_analysis.ipynb -- Notebook with analyses to understand the nature COVID-19 rumours and how they differ to their non-rumour counterpart
  • dataloader.py -- Source code shared by multiple notebooks for loading and processing Twitter data
• submissions/ -- Submissions to the COMP90042 CodaLab competition (Link to competition)

For further information, please refer to the project report attached to this submission.