A graph representation learning algorithm extracts the node embeddings of a complex network from the information cascades. Developed from scratch by C++ and OpenMPI. (The MPI version will be released soon)
The input file starts with NodeID,NodeName per line, followed by an empty line and then each cascade per line
CascadeID;NodeID1,Timestamp1,NodeID2,Timestamp2
where NodeID and CascadeID need to be integers. For example, a valid input file looks like
1,1
2,2
3,3
1;1,0.1,2,0.2,3,0.3
Compile the code with OpenMPI support and execute:
mpirun -n=2 ./sample input=<path to input file> d=10 max_iterations=20
Our paper "Scalable Prediction of Global Online Media News Virality" is published at IEEE Transactions on Computational Social Systems.
@article{lu2018scalable,
title={Scalable Prediction of Global Online Media News Virality},
author={Lu, Xiaoyan and Szymanski, Boleslaw K},
journal={IEEE Transactions on Computational Social Systems},
number={99},
pages={1--13},
year={2018},
publisher={IEEE}
}