In short, a library for computing dynamic pagerank. See [1] and [2] for further details.

1. Ryan Rossi and David Gleich: Dynamic PageRank using Evolving Teleportation, Algorithms and Models for the Web Graph, vol. 7323 of LNCS, pages 126-137. Springer, 2012.

2. David F. Gleich, Ryan A. Rossi, A Dynamical System for PageRank with Time-Dependent Teleportation, Internet Mathematics, 10:1-2, 188-217, 2014.

These codes are research prototypes and may not work for you. No promises. But do email if you run into problems.

• dynamic_pagerank.zip

Unzip data into dynamic_pagerank directory

• dynamic-pr-data.zip
• dynamic-pr-data.7z

Start matlab in the directory where you unzipped the dynamic_pagerank.zip file

$ matlab
>> setup_paths
>> load('data/wiki-24hours');

This should work on Mac OSX (Lion tested) and Ubuntu linux (10.10 tested) with Matlab R2011a.

>> v = normcols(v);
>> X = dynamic_pagerank(A,v);

See examples.m for additional examples

Please let us know if you run into any issues.

The package is organized by directory

/ : All of the main matlab codes (dynamic_pagerank.m,...)

ranking : dynamic ranking codes and figures

forecasting : simple models for prediction using Dynamic PageRank

clustering : experimental codes for identifying trends and similar vertices

causality : codes for computing Granger causality between vertices

data : graphs, precomputed data, and script files for extracting and parsing page views

web : this information and all the figures