Does it cost anything?
No. PEGASUS is an open source software under Apache License 2.0.
If you use PEGASUS for research or commercial purposes, please let us know
your institution(company) and whether it's ok to mention it among the users of PEGASUS.
How do I prepare graph?
PEGASUS works on graphs with TAB-separated plain text format. Each line contains the source and destination
node id of an edge. The node id starts from 0. For example, here is an example graph 'catepillar_star.edge' which is
included in the PEGASUS installation file. It has 16 nodes.
How do I setup a Hadoop cluster?
In most cases, you will use a Hadoop cluster set up by someone else, and thus you don't need to set up a Hadoop
cluster yourself. However, you might want to set up a Hadoop cluster for yourself for some occasions. In that case,
there are several useful documents:
Hadoop Cluster Setup
Cloudera Hadoop Distribution
Note that Hadoop works on a single machine also. Thus, you can run Hadoop on your laptop, for example, and run
PEGASUS on it. For instructions of setting up Hadoop on a single machine, refer to the following web page:
Hadoop Single-Machine Setup
Who make PEGASUS?
PEGASUS is researched and developed by the CMU data mining group led by Prof. Christos Faloutsos. The main
U Kang: Algorithm Design, and Implementation
Duen Horng Chau: Homepage, and Algorithm Design
Christos Faloutsos: Algorithm Design
What are the main advantages of PEGASUS?
The most important advantage of PEGASUS is the scalability. PEGASUS provides graph mining algorithms for
billion-scale graphs which are serveral orders of magnitude larger than previous works. Other advantages include
being an open-source, and faster running time because of the parallel algorithm on top of cloud computing
What is the future of PEGASUS?
PEGASUS will be extended in two directions. At the application level, PEGASUS will incorporate more data mining
and machine learning algorithms including eigenvalue computation and inferences of graphical models. At the data
magamenet level, PEGASUS will be equipped with efficient indexing methods for graphs.
SCHOOL OF COMPUTER SCIENCE