Graph Compression algorithms and software

Summer 2015
With Panagiotis Liakos we have improved the graph representations in popular graph processing systems. More details to be added soon.

Summer 2014
With Panagiotis Liakos and Mike Sioutis we have improved the state-of-the-art algorithm for compressing web and social network graphs.
The algorithm is proper for compressing graphs created by human activity (e.g., web/social network/citation graphs), or any other sparse graph exhibiting the locality of reference property.
Our approach is outlined in this presentation (presented at CIKM '14). The source code, written in Java, can be found here.

Spring 2013
With Panagiotis Liakos and Mike Sioutis we have developed a graph compression algorithm that allows efficient mining of the graph's elements.
The algorithm is proper for compressing graphs created by human activity (e.g., web/social network/citation graphs), or any other sparse graph exhibiting the locality of reference property.
Our approach is outlined in this poster (presented at WSDM '13 1st place in data challenge). The source code, written in Python, can be found here.

More details to be added soon.