Code for our KDD'2015 paper: "Influence at Scale: Distributed Computation of Complex Contagion in Networks" with Yaron Singer and Brendan Lucier.
The code implements a sampling-based algorithm for algorithm, that allows for efficient parallel implementation in distributed frameworks (e.g., Hadoop, Spark). The algorithm a more scalable solution to estimating the influence function for the Independent Cascade influence diffusion process.
Note: the uploaded current code is partial. Some of the Hadoop code is currently under review and should be uploaded soon.