Requirements

python
PyTorch
pandas
tqdm
numpy
scikit_learn
matploblib
numba

Dataset

Download the public data

 MOOC/Reddit(http://snap.stanford.edu/jodie) 
 Amazon(http://jmcauley.ucsd.edu/data/amazon/)

Preprocess the data

 graph file:  `***.csv` 
 	source id, destination id, time, label 

 addtional attributes: `ml_***.npy` 

Running the experiments

Run graph augmentation

python generateNeg.py --data reddit 	

Run anomaly detection

python main.py -d dataset --g_neg False/True --n_degree history_length partner_size --mask mask_num_source mask_num_desti