Subgatt:
This is a tensorflow based implementation of Subgraph Attention as discussed in the paper.
Dataset:
- The dataset_graph folder contains all the datasets which we used in experiments of graph classification.
How to run:
- For Graph Classification: (Default dataset is set to MUTAG) python graphclassification.py
Requirements:
- python (version 3.6 or above)
- tensorflow (version 1.13)
- networkx
- keras
- numpy
- pickle
- scipy
- pandas
- collections
Parameters:
-
For Graph Classification: dataset: The name of the dataset
epoch: Number of epochs to train the mdoel; sub_samp: Number of subgraph samples for each node; sub_leng: The maximum length of any subgraph; pool_rt: Pooling ratio; pool_lay: Number of SubGattPool layers; sub_lay: Number of SubGatt attention layer; learning_rate: Learning rate; embd_dim: Embedding dimension
We can specify these parameters while running these python files. For eg: To specify any other dataset, run following command: python graph_classification.py --dataset NCI1