End-to-end learning from complex multigraphs with latent-graph convolutional networks.
See requirements.txt
.
Zipped data files in pickle format can be found in the data folder. Testing can be done with files followed by a _500
prefix. These contain less than 500 nodes and around 1250 transactions.
Model code can be found in models/LGCN.py
.
Example notebook can be found in the main directory.