%0 Conference Paper %1 kling2014detecting %A Kling, Christoph Carl %A Kunegis, Jérôme %A Sizov, Sergej %A Staab, Steffen %B Proceedings of the 7th ACM international conference on Web search and data mining %D 2014 %K %P 603--612 %T Detecting non-gaussian geographical topics in tagged photo collections %U https://userpages.uni-koblenz.de/~kunegis/paper/kling-multi-dirichlet-geographical-topic-model.pdf %X Nowadays, large collections of photos are tagged with GPS coordinates. The modelling of such large geo-tagged corpora is an important problem in data mining and information retrieval, and involves the use of geographical information to detect topics with a spatial component. In this paper, we propose a novel geographical topic model which captures dependencies between geographical regions to support the detection of topics with complex, non-Gaussian distributed spatial structures. The model is based on a multi-Dirichlet process (MDP), a novel generalisation of the hierarchical Dirichlet process extended to support multiple base distributions. Our method thus is called the MDP-based geographical topic model (MGTM). We show how to use a MDP to dynamically smooth topic distributions between groups of spatially adjacent documents. In systematic quantitative and qualitative evaluations using independent datasets from prior related work, we show that such a model can exploit the adjacency of regions and leads to a significant improvement in the quality of topics compared to the state of the art in geographical topic modelling.

@inproceedings{kling2014detecting, abstract = {Nowadays, large collections of photos are tagged with GPS coordinates. The modelling of such large geo-tagged corpora is an important problem in data mining and information retrieval, and involves the use of geographical information to detect topics with a spatial component. In this paper, we propose a novel geographical topic model which captures dependencies between geographical regions to support the detection of topics with complex, non-Gaussian distributed spatial structures. The model is based on a multi-Dirichlet process (MDP), a novel generalisation of the hierarchical Dirichlet process extended to support multiple base distributions. Our method thus is called the MDP-based geographical topic model (MGTM). We show how to use a MDP to dynamically smooth topic distributions between groups of spatially adjacent documents. In systematic quantitative and qualitative evaluations using independent datasets from prior related work, we show that such a model can exploit the adjacency of regions and leads to a significant improvement in the quality of topics compared to the state of the art in geographical topic modelling.}, added-at = {2016-08-01T20:56:01.000+0200}, author = {Kling, Christoph Carl and Kunegis, J{\'e}r{\^o}me and Sizov, Sergej and Staab, Steffen}, biburl = {https://www.bibsonomy.org/bibtex/2405d1289faa3284bd69a01d004e4950e/nosebrain}, booktitle = {Proceedings of the 7th ACM international conference on Web search and data mining}, interhash = {bc1e3707c6d3fcf47cca64b0ab92274a}, intrahash = {405d1289faa3284bd69a01d004e4950e}, keywords = {}, organization = {ACM}, pages = {603--612}, timestamp = {2016-08-01T20:56:01.000+0200}, title = {Detecting non-gaussian geographical topics in tagged photo collections}, url = {https://userpages.uni-koblenz.de/~kunegis/paper/kling-multi-dirichlet-geographical-topic-model.pdf}, year = 2014 }