Topic and Role Discovery in Social Networks with Experiments on Enron and Academic Email Export

@article{citeulike:2296417, abstract = {Previous work in social network analysis (SNA) has modeled the existence of links from one entity to another, but not the attributes such as language content or topics on those links. We present the Author-Recipient-Topic (ART) model for social network analysis, which learns topic distributions based on the direction-sensitive messages sent between entities. The model builds on Latent Dirichlet Allocation (LDA) and the Author-Topic (AT) model, adding the key attribute that distribution over topics is conditioned distinctly on both the sender and recipient---steering the discovery of topics according to the relationships between people. We give results on both the Enron email corpus and a researcher's email archive, providing evidence not only that clearly relevant topics are discovered, but that the ART model better predicts people's roles and gives lower perplexity on previously unseen messages. We also present the Role-Author-Recipient-Topic (RART) model, an extension to ART that explicitly represents people's roles.}, author = {Mccallum, Andrew and Wang, Xuerui and Corrada-Emmanuel, Andres}, citeulike-article-id = {2296417}, citeulike-linkout-0 = {http://www.jair.org/papers/paper2229.html}, journal = {Journal of Artificial Intelligence Research}, keywords = {discovery, language-model, lda, role, topic}, pages = {249--272}, posted-at = {2008-01-28 05:24:41}, priority = {2}, title = {Topic and Role Discovery in Social Networks with Experiments on Enron and Academic Email}, url = {http://www.jair.org/papers/paper2229.html}, volume = {30}, year = {2007} }

TY - JOUR ID - citeulike:2296417 L3 - citeulike-article-id:2296417 N2 - Previous work in social network analysis (SNA) has modeled the existence of links from one entity to another, but not the attributes such as language content or topics on those links. We present the Author-Recipient-Topic (ART) model for social network analysis, which learns topic distributions based on the direction-sensitive messages sent between entities. The model builds on Latent Dirichlet Allocation (LDA) and the Author-Topic (AT) model, adding the key attribute that distribution over topics is conditioned distinctly on both the sender and recipient---steering the discovery of topics according to the relationships between people. We give results on both the Enron email corpus and a researcher's email archive, providing evidence not only that clearly relevant topics are discovered, but that the ART model better predicts people's roles and gives lower perplexity on previously unseen messages. We also present the Role-Author-Recipient-Topic (RART) model, an extension to ART that explicitly represents people's roles. JF - Journal of Artificial Intelligence Research EP - 272 TI - Topic and Role Discovery in Social Networks with Experiments on Enron and Academic Email VL - 30 SP - 249 KW - discovery KW - language-model KW - lda KW - role KW - topic AU - Mccallum, Andrew AU - Wang, Xuerui AU - Corrada-Emmanuel, Andres PY - 2007/// UR - http://www.jair.org/papers/paper2229.html ER -