On Discovering Moving Clusters in Spatio-temporal Data

Abstract

Abstract.A moving cluster is defined by a set of objects that move close to each other for a long time interval. Real-life examples are a group of migrating animals, a convoy of cars moving in a city, etc. We study the discovery of moving clusters in a database of object trajectories. The difference of this problem compared to clustering trajectories and mining movement patterns is that the identity of a moving cluster remains unchanged while its location and content may change over time. For example, while a group of animals are migrating, some animals may leave the group or new animals may enter it. We provide a formal definition for moving clusters and describe three algorithms for their automatic discovery: (i) a straight-forward method based on the definition, (ii) a more efficient method which avoids redundant checks and (iii) an approximate algorithm which trades accuracy for speed by borrowing ideas from the MPEG-2 video encoding. The experimental results demonstrate the efficiency of our techniques and their applicability to large spatio-temporal datasets.

Modeling Herds and Their Evolvements from Trajectory Data

“…A method to calculate the longest duration of flocks and meetings in spatial-temporal data has been proposed [10]. Definition of moving clusters and efficient algorithms to identify them have been proposed [14]. A recent work tries to find a set of individual trajectories that share the property of visiting the same sequence of places with similar travel times [9], which is related but not the focus of this paper.…”

“…DBSCAN [7], is attractive when one needs to identify arbitrary shaped clusters. In some applications, the location and content of spatial clusters may change over time, which requires a formal definition for moving clusters and algorithms for automatic discoveries [14].…”

Moving objects: Combining gradual rules and spatio-temporal patterns

“…A moving object cluster can be defined in both spatial and temporal dimensions: (1) a group of moving objects should be geometrically close to each other, (2) they should be together for at least some minimum numbers of certain timestamps. In this context, many recent studies are interested in mining moving object clusters including moving clusters [4], flocks [6,7,13], convoys [1,2,12] and trajectory [5,11,3,14].…”

“…For instance, let take a look at Fig 1, there are 12 objects moving independently in 10 minutes. Individual moving objects clusters techniques including moving cluster [4], flock [6,7], convoy [1,2,12] and trajectory [5,11,3] detecting are not efficient in this situation. In fact, these techniques require objects to be together for at least some minimum numbers of certain timestamps; therefore, they are not adapted to extract some interesting patterns in this context.…”

“…A recent study by Jeung [1,2] proposes the convoy concept, an extension of flock, where spatial clustering is based on density. In [4], authors propose to extract moving clusters, i.e. a group of moving objects which have considerably portion of overlap at any two consecutive timestamps.…”