The Topic Modelling research repository for the Digital Circular Electrochemical Economy (DCEE) project at Heriot-Watt University. This research is funded by Digital Circular Electrochemical Economy (EP/V042432/1), and the UK Research and Innovation (UKRI) Interdisciplinary Centre for Circular Chemical Economy (EP/V011863/1). In response to this call, we have united a cross-disciplinary team of leading researchers from three UK universities: Imperial College London, Loughborough University, and Heriot-Watt University.
The datasets and experimental results will be made publicly available following the EPSRC Data Storage Policy and GDPR Regulations. Currently, only the code for the models, hyperparameter optimisation experiments, and data preprocessing scripts are publicly available. Full datasets and results will be available after approval.
This research has been submitted to the JCR Q1 Elsevier journal "Energy and AI". The preprint is now available on arXiv.
To ensure compatibility with the code, it is recommended to create a Python 3.8 virtual environment. Follow these steps:
- Install Python 3.8 and virtualenv if you haven't already.
- Create a virtual environment:
virtualenv -p python3.8 venv
- Activate the virtual environment:
- On Windows:
venv\Scripts\activate
- On Unix or MacOS:
source venv/bin/activate
- On Windows:
- Install the required packages:
pip install -r requirements.txt
- Install Anaconda or Miniconda if you haven't already.
- Create a conda environment with Python 3.8:
conda create --name dcee python=3.8
- Activate the conda environment:
conda activate dcee
- Install the required packages:
pip install -r requirements.txt
The repository contains scripts for different models (BERTopic, CorEx, LDA) and preprocessing steps. You can find the scripts in the scripts
directory. Each subdirectory contains Jupyter notebooks (.ipynb
) and Python scripts (.py
) for hyperparameter tuning and grid search operations.
To run a specific script, navigate to its directory and execute the script. For example:
cd scripts/bertopic
python bert_grid_guardian.py
This project is licensed under the Apache 2.0 License - see the LICENSE file for details.
For any questions or further information, please contact the project team at dcee.org.uk.
@article{song2024navigating,
title={Navigating Public Sentiment in the Circular Economy through Topic Modelling and Hyperparameter Optimisation},
author={Song, Junhao and Yuan, Yingfang and Chang, Kaiwen and Xu, Bing and Xuan, Jin and Pang, Wei},
journal={arXiv preprint arXiv:2405.10452},
year={2024}
}