Tractable Group Detection on Large Link Data Sets Export

@inproceedings{citeulike:580809, abstract = {Discovering underlying structure from co-occurrence data is an important task in a variety of fields, including: insurance, intelligence, criminal investigation, epidemiology, human resources, and marketing. Previously Kubica et. al. presented the group detection algorithm (GDA) - an algorithm for finding underlying groupings of entities from co-occurrence data. This algorithm is based on a probabilistic generative model and produces coherent groups that are consistent with prior knowledge. Unfortunately, the optimization used in GDA is slow, potentially making it infeasible for many large data sets. To this end, we present k-groups - an algorithm that uses an approach similar to that of k-means to significantly accelerate the discovery of groups while retaining GDA's probabilistic model. We compare the performance of GDA and k-groups on a variety of data, showing that k-groups' sacrifice in solution quality is significantly offset by its increase in speed.}, address = {Washington, DC, USA}, author = {Kubica, Jeremy and Moore, Andrew and Schneider, Jeff}, booktitle = {ICDM '03: Proceedings of the Third IEEE International Conference on Data Mining}, citeulike-article-id = {580809}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=951949.952107}, citeulike-linkout-1 = {http://www.autonlab.org/autonweb/14635/version/2/part/5/data/kubicaKgroups.pdf?branch=main&language=en}, citeulike-linkout-2 = {http://citeseer.ist.psu.edu/640879}, editor = {Wu, Xindong and Tuzhilin, Alex and Shavlik, Jude}, isbn = {0769519784}, keywords = {community, groupmodel, topicinference}, month = {November}, pages = {573--576}, posted-at = {2006-04-09 13:33:01}, priority = {0}, publisher = {IEEE Computer Society}, title = {Tractable Group Detection on Large Link Data Sets}, url = {http://portal.acm.org/citation.cfm?id=951949.952107}, year = {2003} }

TY - CONF ID - citeulike:580809 L3 - citeulike-article-id:580809 N2 - Discovering underlying structure from co-occurrence data is an important task in a variety of fields, including: insurance, intelligence, criminal investigation, epidemiology, human resources, and marketing. Previously Kubica et. al. presented the group detection algorithm (GDA) - an algorithm for finding underlying groupings of entities from co-occurrence data. This algorithm is based on a probabilistic generative model and produces coherent groups that are consistent with prior knowledge. Unfortunately, the optimization used in GDA is slow, potentially making it infeasible for many large data sets. To this end, we present k-groups - an algorithm that uses an approach similar to that of k-means to significantly accelerate the discovery of groups while retaining GDA's probabilistic model. We compare the performance of GDA and k-groups on a variety of data, showing that k-groups' sacrifice in solution quality is significantly offset by its increase in speed. SN - 0769519784 AD - Washington, DC, USA EP - 576 TI - Tractable Group Detection on Large Link Data Sets PB - IEEE Computer Society T2 - ICDM '03: Proceedings of the Third IEEE International Conference on Data Mining SP - 573 KW - community KW - groupmodel KW - topicinference AU - Kubica, Jeremy AU - Moore, Andrew AU - Schneider, Jeff A2 - Wu, Xindong A2 - Tuzhilin, Alex A2 - Shavlik, Jude PY - 2003/11// UR - http://portal.acm.org/citation.cfm?id=951949.952107 ER -