Deep Dynamic Relational Classifiers (DDRC)
This is a python implementation of Deep Dynamic Relational Classifiers (DDRC) for the task of (semi-supervised) classification of nodes in a graph, as described in our paper:
Hogun Park, John Moore, Jennifer Neville, Deep Dynamic Relational Classifiers: Exploiting Dynamic Neighborhoods in Complex Networks, Proc. of MAISoN Workshop in 10th ACM Conference on Web Search and Data Mining (WSDM 2017), 2017.
Usage
Example Usage
python main/arg_rnn_classifier_with_attr.py --dataset dblp --hiddenunits 256 --timedistunits 64 --maxpooling 4 --randoption 0 --dntypeoption 0 --epochs 500 --nclasses 6 --preprocesseddataset 0 --parallel 1
Requirements
- numpy
- scipy
- keras
- theano
(may have to be independently installed)