Inferring Networks of Diffusion and Influence

ConNIe (Convex Network Inference) infers hidden/latent social networks from cascade data. The algorithm is described in detail in the paper:

On the Convexity of Latent Social Network Inference by S. A. Myers, J. Leskovec. Neural Information Processing Systems (NIPS), 2010.

ConNIe uses SNOPT7, developed by the Systems Optimization Lab here at Stanford, which is a Fortran-based nonlinear optimization solver. We cannot freely distribute the solver here, but if you are able to obtain a licence (for example, faculty and students at Stanford can obtain it for free), then you can use the original scalable implementation of ConNIe.

Download ConNIe

To use, simply unzip the directory and call the "connie" method from MATLAB's command line. Assuming that SNOPT7 (including its MATLAB interface) has been installed, there is no additional installation required. See the comments at the begining of connie.m for an explanation of how to run the code.

If you have suggestions, comments, or problems with the code, please contact [username]AT[domain].edu where [username] is "samyers" and [domain] is "stanford."