We present a framework for the partitioning of a spatial trajectory in a sequence of segments based on spatial density and temporal criteria. The result is a set of temporally separated clusters interleaved by sub-sequences of unclustered points. A major novelty is the proposal of an outlier or noise model based on the distinction between intra-cluster (local noise) and inter-cluster noise (transition): the local noise models the temporary absence from a residence while the transition the definitive departure towards a next residence. We analyze in detail the properties of the model and present a comprehensive solution for the extraction of temporally ordered clusters. The effectiveness of the solution is evaluated first qualitatively and next quantitatively by contrasting the segmentation with ground truth. The ground truth consists of a set of trajectories of labeled points simulating animal movement. Moreover, we show that the approach can streamline the discovery of additional derived patterns, by presenting a novel technique for the analysis of periodic movement. From a methodological perspective, a valuable aspect of this research is that it combines the theoretical investigation with the application and external validation of the segmentation framework. This paves the way to an effective deployment of the solution in broad and challenging fields such as e-science.

No abstract

We present an approach to the discovery and characterization of relevant locations and related mobility patterns in symbolic trajectories built on call detail records-CDRs-of mobile phones (telco trajectories). While the discovery of relevant locations has been widely investigated for continuous spatial trajectories (e.g., stay points detection methods), it is not clear how to deal with the problem when the movement is defined over a discrete space and the locations are symbolic, noisy and irregularly sampled, such as in telco trajectories. In this paper, we propose a methodological approach structured in two steps, called trajectory summarization and summary trajectories analysis, respectively, the former for removing noise and irrelevant locations; the latter to synthesize key mobility features in a few novel indicators. We evaluate the methodology over a dataset of approx 17,000 trajectories with 55 million points and spanning a period of 67 days. We find that trajectory summarization does not compromise data utility, while significantly reducing data size. Moreover, the mobility indicators provide novel insights into human mobility behavior.

Segmentation techniques partition a sequence of data points in a series of disjoint sub-sequences -segments -based on some criteria. Depending on the context and the nature of data points, segments can be given an approximated representation. The final result is a summarized representation of the sequence. This intuitive mechanism has been extensively studied, for example, for the summarization of time series in order to preserve the 'shape' of the sequence while omitting irrelevant details. This survey focuses on the use of segmentation methods for extracting behavioral information from individual mobility data, in particular from spatial trajectories. Such information can then be given a compact representation in the form of summarized trajectories, e.g., semantic trajectories, symbolic trajectories. Two major streams of research are discussed, spanning computational geometry and data mining respectively, that are emblematic of the multiplicity of views.

No abstract