handasontam/EdgeProp

Dependencies

$ pip install -r requirements.txt

# follow https://docs.dgl.ai/en/0.4.x/install/ to install dgl with gpu
# e.g.
$ pip install dgl-cu101

Run experiment (EdgeProp)

# Copy features.csv, labels.csv and network.csv to the data directory
# preprocess the dataa
$ python preprocess.py --data-dir ./data
# Start the experiment (120 experiments in total)
$ python run.py

Run experiment (Classical Baselines)

$ python main.py --data-dir ./data --model-dir experiments/classical_baseline/ --gpu 0

network.csv:

srcId,dstId,timestamp(YYYYMMDDMISS),money,type
...

features.csv

nodeId, age, count, ...
...

labels.txt

nodeId,label
...