EvoGraph [SIGKDD18]

EvoGraph: An Effective and Efficient Graph Upscaling Method for Preserving Graph Properties

Usage (Simple example)

• Download source codes by git clone https://github.com/chan150/EvoGraph.git

• Compile ./compile.sh (for distributed environment) or ./compile-local.sh (for a single machine)

• Execute EvoGraph in a single machine ./run-local.sh output -gs.input toy -gs.sf 2

• Execute EvoGraph in distributed machines ./run.sh output -gs.input toy -gs.sf 2

Usage (Detail example)

• upscaling a toy graph with a scale factor 2 by using 120 machines/threads ./run.sh output -gs.input toy -gs.sf 2 -m 120

Parameters

• -gs.input input path in master computer
• -gs.sf scale factor
• -machine or -m number of machines/threads (e.g. -m 120)

Parameters (optional for a large input file)

• -gs.eid number of edges in original graph (optional; e.g. -gs.eid 10427)
• -gs.vid number of vertices in original graph (optional; e.g. -gs.vid 1024)

License

Licensed under the Apache License, Version 2.0.

Citing the paper

For more details about EvoGraph, please refer to our paper. You can download the paper via open access in my github.io site.

If you use EvoGraph, please cite:

@inproceedings{Park:2018:EEE:3219819.3220123,
 author = {Park, Himchan and Kim, Min-Soo},
 title = {EvoGraph: An Effective and Efficient Graph Upscaling Method for Preserving Graph Properties},
 booktitle = {Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery \&\#38; Data Mining},
 series = {KDD '18},
 year = {2018},
 isbn = {978-1-4503-5552-0},
 location = {London, United Kingdom},
 pages = {2051--2059},
 numpages = {9},
 url = {http://doi.acm.org/10.1145/3219819.3220123},
 doi = {10.1145/3219819.3220123},
 acmid = {3220123},
 publisher = {ACM},
 address = {New York, NY, USA},
 keywords = {barabasi-albert, graph generation, graph upscaling, parallel computation, preferential attachment},
}