10.7910/DVN/GGHMFT Krafft, Peter Peter Krafft Massachusetts Institute of Technology Moore, Juston Juston Moore University of Massachusetts Amherst Desmarais, Bruce Bruce Desmarais University of Massachusetts Amherst Wallach, Hanna Hanna Wallach University of Massachusetts Amherst Harvard Dataverse 2012 network analysis, topic modeling, machine learning, political science, latent space network analysis, topic modeling, machine learning, political science Bruce Desmarais Bruce Desmarais University of Massachusetts Amherst Bruce Desmarais Bruce Desmarais Bruce Desmarais Bruce Desmarais University of Massachusetts Amherst 2012 2012 2012-12-13 2012-12-13 2012-12-13 government email archive 1902.1/19575 /papers/files/nips25/NIPS2012_1288.pdf 10.7910/DVN/GGHMFT/3ZBIOB 10.7910/DVN/GGHMFT/XA1EYF 10.7910/DVN/GGHMFT/T4SUPM 10.7910/DVN/GGHMFT/MEPLGL 10.7910/DVN/GGHMFT/17LTHC 3015 50883 197019 21913012 641 text/plain; charset=US-ASCII text/plain; charset=US-ASCII text/plain; charset=US-ASCII text/plain; charset=US-ASCII text/plain; charset=US-ASCII 2.0 Creative Commons CC0 1.0 Universal Public Domain Dedication. We introduce a joint model of network content and context designed for exploratory analysis of email networks via visualization of topic-specific communication patterns. Our model is an admixture model for text and network attributes which uses multinomial distributions over words as mixture components for explaining text and latent Euclidean positions of actors as mixture components for explaining network attributes. We validate the appropriateness of our model by achieving state-of-the-art performance on a link prediction task and by achieving semantic coherence equivalent to that of latent Dirichlet allocation. We demonstrate the capability of our model for descriptive, explanatory, and exploratory analysis by investigating the inferred topic-specific communication patterns of a new government email dataset, the New Hanover County email corpus. This work was supported in part by the Center for Intelligent Information Retrieval and in part by the NSF GRFP under grant #1122374. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect those of the sponsors. United States New Hanover County, North Carolina,