[TOIS] Adaptive Sampling-based Dynamic Graph Learning for Information Diffusion Prediction

This is the implementation of ASDIP: [Adaptive Sampling-based Dynamic Graph Learning for Information Diffusion Prediction].

Requirements

• python == 3.9.18
• pytorch == 2.0.1
• numba == 0.55.1
• numpy == 1.21.6
• pandas == 2.0.3
• scikit-learn == 1.3.0
• dgl == 1.1.2
• tqdm
• wandb

Dataset

• Download the preprocessed dataset from Baidu Yun （extract code nihb）
• create a directory ./data/dynamic and put the downloaded dataset into the directory.

Run

Create the directories to store the running results

mkdir log results saved_models

Running command

#Memetracker
python train_ASDIP.py --prefix std --dataset douban --gpu 0 --causal before after none
#Douban
python train_ASDIP.py --prefix std --dataset memetracker --gpu 0 --causal before after none
#Weibo
python train_ASDIP.py --prefix std --dataset weibo --gpu 0 --causal before after none