This repository contains models for fake news detection of COVID-19 data with spatial and temporal features. Jupyter notebooks found here contain of the following:
- LSTM Model without spatial and temporal information
- LSTM Model with spatial and temporal information
- Hyperparameter optimization of the two models
Install the dependencies from the requirements.txt file
pip install -r requirements.txt
After installing the required dependencies, the project files are LSTM_RNN_Implementation_with_hyperparameterisation.ipynb and Country_Date.ipynb, where models with and without spatial and temporal features are present respectively. These two notebooks were run on Google Colab to use their GPU for faster training performance; The required dataset is present under the Datasets folder, which in our case had to be uploaded to drive, and the folder mounted on drive.
Hyperparameter Optimization was done with the help of a package, talos. The best hyperparametersparameters required were then visualized in the the files Hyperparameter_Optimization1.ipynb and Hyperparameter_Optimization2.ipynb.
Autonomio Talos [Computer software]. (2019). Retrieved from http://github.com/ autonomio/talos