- Linux
- Python 3.7
- PyTorch 1.10.0+cu111
Clone this repo.
git clone https://github.com/THUDM/citation-prediction.git
cd citation-prediction
Please install dependencies by
pip install -r requirements.txt
We provide two datasets for author citation prediction. The one is to predict author citations is 2016 and another is to predict author citations in 2022. The two datasets contain different authors. The datasets can be downloaded from BaiduPan with password g5uk or data-2016 & data-2022. Please put the data folder into the project directory.
cd $project_path
export CUDA_VISIBLE_DEVICES='?' # specify which GPU(s) to be used
# processing: set pred_year = 2016/2022 in process.py
python process.py
# ARIMA: set pred_year = 2016/2022 below __main__ function
python arima.py
# regressor: set pred_year = 2016/2022 below __main__ function
python regressor.py
# LSTM: set pred_year = 2016/2022 below __main__ function
python lstm.py
# EvolveGCN
cd evolvegcn
python run_exp_inf.py --config_file ./experiments/parameters_inf_2016.yaml
python run_exp_inf.py --config_file ./experiments/parameters_inf_2022.yaml
Evaluation metrics: RSME
2016 | 2022 | |
---|---|---|
ARIMA | 1225 | 23920 |
LR | 562 | 22057 |
GBRT | 553 | 21777 |
LSTM | 1034 | 25409 |
EvolveGCN | 969 | 22841 |