TY - CONF TI - Rethinking LDA: Why Priors Matter AU - Wallach, Hanna M. AU - Mimno, David AU - McCallum, Andrew T3 - NIPS'09 AB - Implementations of topic models typically use symmetric Dirichlet priors with fixed concentration parameters, with the implicit assumption that such "smoothing parameters" have little practical effect. This paper explores several classes of structured priors for topic models. C1 - USA C3 - Proceedings of the 22Nd International Conference on Neural Information Processing Systems DA - 2009/// PY - 2009 SP - 1973 EP - 1981 LA - en PB - Curran Associates Inc. SN - 978-1-61567-911-9 ST - Rethinking LDA UR - http://dl.acm.org/citation.cfm?id=2984093.2984314 Y2 - 2019/01/14/00:00:00 KW - Topic model algorithm KW - Topic modeling ER -