Scientists study trajectory data to understand trends in movement patterns, such as human mobility for traffic analysis and urban planning. In this paper, we introduce a novel trajectory clustering technique whose central idea is to use vector fields to induce a notion ofsimilarity between trajectories, letting the vector fields themselves define and represent each cluster. We present an efficient algorithm to find a locally optimal clustering oftrajectories into vector fields, and demonstrate how vector-field k-means can find patterns missed by previous methods. We present experimental evidence ofits effectiveness and efficiency using several datasets, including historical hurricane data, GPS tracks ofpeople and vehicles, and anonymous cellular radio handoffs from a large service provider.
%0 Journal Article
%1 ferreira2013vector
%A Ferreira, Nivan
%A Klosovski, James T.
%A Scheidegger, Carlos E.
%A Silva, Claudio T.
%D 2013
%J Eurographics Conference on Visualization (EuroVis) 2013
%K Computer Forum Graphics
%N 3
%T Vector Field k-Means: Clustering Trajectories By Fitting Multiple Vector Fields
%V 32
%X Scientists study trajectory data to understand trends in movement patterns, such as human mobility for traffic analysis and urban planning. In this paper, we introduce a novel trajectory clustering technique whose central idea is to use vector fields to induce a notion ofsimilarity between trajectories, letting the vector fields themselves define and represent each cluster. We present an efficient algorithm to find a locally optimal clustering oftrajectories into vector fields, and demonstrate how vector-field k-means can find patterns missed by previous methods. We present experimental evidence ofits effectiveness and efficiency using several datasets, including historical hurricane data, GPS tracks ofpeople and vehicles, and anonymous cellular radio handoffs from a large service provider.
@article{ferreira2013vector,
abstract = {Scientists study trajectory data to understand trends in movement patterns, such as human mobility for traffic analysis and urban planning. In this paper, we introduce a novel trajectory clustering technique whose central idea is to use vector fields to induce a notion ofsimilarity between trajectories, letting the vector fields themselves define and represent each cluster. We present an efficient algorithm to find a locally optimal clustering oftrajectories into vector fields, and demonstrate how vector-field k-means can find patterns missed by previous methods. We present experimental evidence ofits effectiveness and efficiency using several datasets, including historical hurricane data, GPS tracks ofpeople and vehicles, and anonymous cellular radio handoffs from a large service provider.},
added-at = {2018-07-01T13:42:06.000+0200},
author = {Ferreira, Nivan and Klosovski, James T. and Scheidegger, Carlos E. and Silva, Claudio T.},
biburl = {https://puma.ub.uni-stuttgart.de/bibtex/2670db3a5238d0d631dfc280f607790f5/mwigger},
interhash = {355b25028750841adada3b2de164caf3},
intrahash = {670db3a5238d0d631dfc280f607790f5},
journal = {Eurographics Conference on Visualization (EuroVis) 2013},
keywords = {Computer Forum Graphics},
number = 3,
timestamp = {2018-09-14T09:59:55.000+0200},
title = {Vector Field k-Means: Clustering Trajectories By Fitting Multiple Vector Fields},
volume = 32,
year = 2013
}