@techreport{TD:100841,
	att_abstract={{Scientists have often studied trajectory data to better understand trends in movement patterns, including the migration of animals, storm track predictions, as well as human mobility for traffic analysis and urban planning.  With tracking devices becoming smaller and more prevalent, the amount of trajectory data being gathered is steadily increasing.  As a result, there is a tremendous need for scalable and efficient techniques for analyzing this data and discovering the underlying patterns. In this paper, we introduce a novel technique for extracting arbitrary movement patterns from trajectory data which we call Vector Field k-Means. Many traditional trajectory clustering algorithms look for a representative trajectory that best describes each cluster, very much like k-means identifies a representative "center" for each cluster.  Vector field k-means on the other hand, recognizes that in all but the simplest examples, no single trajectory adequately describes a cluster.  Our approach is based on the premise that movement trends in trajectory data can be modeled as flows within multiple vector fields, and the vector field itself is what defines each of the clusters. The use of vector fields to encode movement patterns is natural, as they immediately form a suitable visual representation and we can take advantage of advanced techniques to visualize and analyze them. We demonstrate how vector field k-means can be used to analyze trajectory data and present experimental evidence of its effectiveness and efficiency using several datasets,  including historical hurricane data, GPS tracks of people and vehicles, and anonymous call records from a large communications provider.}},
	att_authors={cs929g, jk140f},
	att_categories={C_IIS.7, C_IIS.2},
	att_copyright={{Blackwell Publishing}},
	att_copyright_notice={{The definitive version was published in  2013. {{, 2013-06-17}}
}},
	att_donotupload={},
	att_private={false},
	att_projects={IV_INFOVIS},
	att_tags={trajectory, clustering, visualization, k-means},
	att_techdoc={true},
	att_techdoc_key={TD:100841},
	att_url={http://web1.research.att.com:81/techdocs_downloads/TD:100841_DS1_2013-09-16T18:14:56.284Z.pdf},
	author={Carlos Scheidegger and James Klosowski and Nivan Ferreira and Claudio Silva},
	institution={{Eurovis 2013}},
	month={June},
	title={{Vector Field k-Means: Detecting Patterns in Trajectory Data by Fitting Multiple Vector Fields}},
	year=2013,
}