Parallel gathering discovery over big trajectory data

Xian, Yongyi; Liu, Yan; Xu, Chuanfei

doi:10.1109/bigdata.2016.7840671

Cited by 9 publications

(6 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We conduct a thorough evaluate of our proposed methods on a real dataset with three algorithms: (1) snapshot based model called gathering method [27], [22]; (2) our trajectory slot model with the point-to-polyline distance measures (referred as TCompanion-P2PL) [23] * 1 ; and (3) the trajectory slot model with the polyline-to-polyline distance measures (referred as TCompanion-PL2PL). We evaluate the following the quality attributes, namely Precision and Recall, Performance and Scalability, Data Shuffling and Intensity and Stability.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Parallel Discovery of Trajectory Companion Pattern and System Evaluation

Xian¹,

Liu²,

Xu³

et al. 2020

Journal of Information Processing

Self Cite

View full text Add to dashboard Cite

Trajectories consist of spatial information of moving objects. Over contious time spans, trajectory data form data streams constantly generated from diverse and geographically distributed sources. Discovery of traveling patterns on trajectory streams such as gathering and companies enables value domain applications. Such a discovery needs to process arrival records in various sources and correlate across records near real-time. Thus techniques for handling trajectory streams should scale on distributed cluster computing. The challenge is at three aspects, namely a data model to represent the continuous trajectory data, the parallelism of the discovery algorithm, and an end-to-end parallel framework. In this paper, we propose a parallel discovery method that consists of 1) a model of partitioning trajectory samples on various time intervals; 2) definition on distance measurements of trajectories; and 3) a parallel discovery algorithm. We build a stream processing workflow and investigate experiments on a public dataset to evaluate the system's performance, scalability, stability, and data intensity. Our method discovers trajectory gathering patterns with low latency and scales as the size of trajectory data grows.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Figure 5 shows the precision, recall and F1-score of our proposed algorithm (TCompanion) and the competitor algorithm (Gathering) [22], [27]. We vary the size of trajectory slot T from 40 to 80 seconds.…”

Section: Precision and Recall Evaluationmentioning

confidence: 99%

Parallel Discovery of Trajectory Companion Pattern and System Evaluation

Xian¹,

Liu²,

Xu³

et al. 2020

Journal of Information Processing

Self Cite

View full text Add to dashboard Cite

show abstract

“…Zhang et al [26] designed a spatiotemporal graph structure to retrieve the maximal gathering group that meets the spatial-temporal constraints. Xian et al [27] formed the gathering pattern discovery framework in a distributed parallel computing system to process large sets of trajectory data. Hung et al [28] designed a new trajectory pattern mining framework-Clustering and Aggregating Clues of Trajectories (CACT).…”

Section: Group Movement Pattern Discoverymentioning

confidence: 99%

Discovery of evolving companion from trajectory data streams

2020

View full text Add to dashboard Cite

The widespread use of position-tracking devices leads to vast volumes of spatial-temporal data aggregated in the form of the trajectory data streams. Extracting useful knowledge from moving object trajectories can benefit many applications, such as traffic monitoring, military surveillance, and weather forecasting. Most of the knowledge gleaned from the trajectory data illustrates different kinds of group patterns, i.e., objects that travel together for some time. In the real world, the trajectory of the moving objects can change with time. Thus, existing approaches can miss a new pattern because they have a stringent requirement for moving object participators in a group movement pattern. To address this issue, we introduced a new type of moving object group pattern called an evolving companion. It allows some members of the group to leave and join anytime if some participators stay connected for all time intervals. In this pattern discovery, we model an incremental discovery solution to retrieve the evolving companion efficiently over the data stream. We evaluated the efficiency and effectiveness of our approach on two real vehicles and one synthetic dataset. Our method performed well compared with existing pattern discovery methods; for example, it was about 50% faster than Tang et al.'s buddy-based clustering method.

show abstract

“…They utilised an improved R‐tree based density clustering algorithm to index moving objects and clusters and a spatio‐temporal graph to retrieve patterns [22, 23]. Xian et al paid attention to both batch and streaming fashion of gathering patterns parallel discovery [24].…”

Section: Related Workmentioning

confidence: 99%

“…In recent years, more and more studies began to process traffic data stream. Besides the above‐mentioned works [7–9, 19, 20, 24], Yu et al studied on a density‐based clustering algorithm for trajectory data stream and tried to discover trajectory clusters in real time [25, 26]. All these related work aimed at processing GPS data stream and provide some foundations for our study.…”

Section: Related Workmentioning

confidence: 99%