%0 Conference Paper %1 citeulike:393513 %A Wang, Xuerui %A Mohanty, Natasha %A Mccallum, Andrew %B KDD Workshop on Link Discovery: Issues, Approaches and Applications (LinkKDD) %D 2005 %K topicinference socialnets community %T Group and Topic Discovery from Relations and Text %U http://www.cs.umass.edu/~mccallum/papers/grouptopic_linkkdd05.pdf %X We present a probabilistic generative model of entity relationships and textual attributes that simultaneously discovers groups among the entities and topics among the corresponding text. Block-models of relationship data have been studied in social network analysis for some time. Here we simultaneously cluster in several modalities at once, incorporating the words associated with certain relationships. Significantly, joint inference allows the discovery of groups to be guided by the emerging topics, and vice-versa. We present experimental results on two large data sets: sixteen years of bills put before the U.S. Senate, comprising their corresponding text and voting records, and 43 years of similar data from the United Nations. We show that in comparison with traditional, separate latent-variable models for words or Blockstructures for votes, the Group-Topic model’s joint inference improves both the groups and topics discovered.

@inproceedings{citeulike:393513, abstract = {We present a probabilistic generative model of entity relationships and textual attributes that simultaneously discovers groups among the entities and topics among the corresponding text. Block-models of relationship data have been studied in social network analysis for some time. Here we simultaneously cluster in several modalities at once, incorporating the words associated with certain relationships. Significantly, joint inference allows the discovery of groups to be guided by the emerging topics, and vice-versa. We present experimental results on two large data sets: sixteen years of bills put before the U.S. Senate, comprising their corresponding text and voting records, and 43 years of similar data from the United Nations. We show that in comparison with traditional, separate latent-variable models for words or Blockstructures for votes, the Group-Topic model’s joint inference improves both the groups and topics discovered.}, added-at = {2006-06-16T10:34:37.000+0200}, author = {Wang, Xuerui and Mohanty, Natasha and Mccallum, Andrew}, biburl = {https://www.bibsonomy.org/bibtex/25cb81f878ffc60ccda8ab45941bcf898/ldietz}, booktitle = {KDD Workshop on Link Discovery: Issues, Approaches and Applications (LinkKDD)}, citeulike-article-id = {393513}, comment = {Extension to elaborate Author-Topic Models and the Stochastik Blockstructure Model. Problem domain: voting in Politics "The Group-Topic Model is a directed graphical model that clusters entities with relations between them, as well as attributes of those relations. The relations may be either directed or undirected and have multiple attributes." They have a rather complex graphical model, to which they apply Gibbs Sampling to get P(group of entity|topic) and P(topic of an event) In fact, they model a single type of relation .oO(is this true?) with a binary attribute, which is the (dis-)agreement of the votes (yes/no/abundant). Their future work: Application of the GT model to scientific citations and allowing mixtures of topics as a relation attribute.}, interhash = {fc58c22171e32fc00ec14e63d33a35dc}, intrahash = {5cb81f878ffc60ccda8ab45941bcf898}, keywords = {topicinference socialnets community}, priority = {0}, timestamp = {2006-06-16T10:34:37.000+0200}, title = {Group and Topic Discovery from Relations and Text}, url = {http://www.cs.umass.edu/~mccallum/papers/grouptopic_linkkdd05.pdf}, year = 2005 }