This is the implementation of CTCP: [Continuous-Time Graph Learning for Cascade Popularity Prediction, IJCAI 2023].
- python == 3.7.0
- pytorch == 1.9.1
- dgl == 0.8.2
- scikit-learn == 1.0
- numpy == 1.21.5
- pandas == 1.1.5
- tqdm == 4.62.3
- Download the preprocessed dataset from Baidu Yun （extract code myu6）
- create a directory
./dataand put the downloaded dataset into the directory.
Create the directories to store the running results
mkdir log results saved_modelsRunning command
#Twitter
python main.py --dataset twitter --prefix std --gpu 0 --epoch 150 --embedding_module aggregate --use_dynamic --use_temporal --use_structural --use_static --dropout 0.6 --predictor merge --lambda 0.1
#APS
python main.py --dataset aps --prefix std --gpu 0 --epoch 150 --embedding_module aggregate --use_dynamic --use_temporal --use_structural --use_static --dropout 0.6 --predictor merge --lambda 0.1
#Weibo
python main.py --dataset weibo --prefix std --gpu 0 --epoch 150 --embedding_module aggregate --use_dynamic --use_temporal --use_structural --use_static --dropout 0.6 --predictor merge --lambda 0.1After running, the log file, results, and trained model are saved under the directories of log, saved_results, and saved_models respectively.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent