4D+SNN: A Spatio-Temporal Density-Based Clustering Approach with 4D Similarity

Oliveira, Ricardo F.; Santos, Maribel Yasmina; Pires, João Moura

doi:10.1109/icdmw.2013.119

Cited by 15 publications

(12 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ricardo [13] et al proposed a new algorithm based on an extension of the SNN (Shared Nearest Neighbor) [14] algorithm, called the 4D+SNN algorithm, which allows the integration of space, time and one or more semantic attributes in the clustering process. This algorithm is able to deal with different data sets and different discovery purposes as the user has the ability to weight the importance of each dimension in the discovery process.…”

Section: Related Workmentioning

confidence: 99%

Novel clustering and analysis techniques for mining spatio-temporal data

Zhang

Eick

2014

Proceedings of the 1st ACM SIGSPATIAL PhD Workshop

View full text Add to dashboard Cite

Due to the advances in remote sensing and sensor networks, different types of dynamic and spatio-temporal datasets become increasingly available. Extracting spatial and temporal patterns from such datasets is very important as it has many applications, such as understanding climate change, geo-targeting, and environment protection. Traditional clustering techniques are inefficient in clustering spatio-temporal data because they do not incorporate the idiosyncrasies of the spatial domain and temporal domain, and do not consider interactions between the two domains. In the presented dissertation research, novel spatiotemporal clustering algorithms and frameworks will be investigated, and post-processing analysis techniques will be developed which visualize and characterize the obtained clusters. The second goal of the dissertation research is to apply the framework to challenging problems of our society while evaluating the proposed algorithms and frameworks and making contributions to their solution. Namely, traffic fatality and air pollution datasets will be used for this purpose. We also present preliminary work in which two spatio-temporal clustering algorithms called ST-SSN and ST-SEP-SSN were evaluated in a case study involving ozone pollution events in the HoustonGalveston-Brazoria (HGB) area, providing a motivation for the planned dissertation research.

show abstract

Section: Related Workmentioning

confidence: 99%

Novel clustering and analysis techniques for mining spatio-temporal data

Zhang

Eick

2014

Proceedings of the 1st ACM SIGSPATIAL PhD Workshop

View full text Add to dashboard Cite

show abstract

“…Ricardo et al [7] proposed a new algorithm based on an extension of the SNN (Shared Nearest Neighbor) [8] algorithm, called the 4D+SNN algorithm, which allows the integration of space, time and semantic attributes into the clustering process. This algorithm is able to deal with different data sets and different discovery purposes as the user has the ability to weight the importance of each dimension in the discovery process.…”

Section: Related Workmentioning

confidence: 99%

A computational framework for finding interestingness hotspots in large spatio-temporal grids

Akdag

Davis

Eick

2014

Proceedings of the 3rd ACM SIGSPATIAL International Workshop on Analytics for Big Geospatial Data

View full text Add to dashboard Cite

Gridded datasets are quite common in scientific computing as many disciplines produce large amounts of samples relying on regular spatial grid-structures that identify locations where measurements are taken. Very large gridded datasets with a lot of variables and measurements pose challenging problems on how to store, query, summarize, visualize and mine such datasets. One interesting and important task when analyzing such datasets is to find interesting contiguous regions, called interestingness hotspots, based on the domain expert's notion of interestingness which is captured in an interestingness function. There has been significant work in using spatio-temporal clustering for mining spatio-temporal datasets in literature. However, in this paper, we present a computational framework which uses an alternative nonclustering approach to obtain interestingness hotspots. We present a novel hotspot growing algorithm which grows interestingness hotspots from seed hotspots, and then post-processes the obtained hotspots to remove or merge overlapping hotspots. We claim that our approach is capable of identifying a much broader class of hotspots, which cannot be identified by traditional distance-based clustering algorithms. Our framework is evaluated in a case study for a very large 4-dimensional gridded air pollution dataset in which we find high correlation hotspots and low variance hotspots with respect to pollutants; moreover, we analyze the scalability of the proposed hotspot growing algorithm.

show abstract

“…Therefore, for TriStar to find ST RoIs it must use some mapping for space and time data into a distance metric. The approach we utilise is a modified version of the equation introduced by Oliveira et al in [10]. Their concept is to normalise the distance between each dimensional entry into the range [0,1] using a maximum dimensional distance.…”

Section: A St Distance Mappingmentioning

confidence: 99%

“…We now present our equation, based on [10], for normalising a ST entry such that we can apply geometric operations to it:…”

Section: A St Distance Mappingmentioning

confidence: 99%

See 1 more Smart Citation

Spatio-Temporal Trajectory Region-of-Interest Mining Using Delaunay Triangulation

Bermingham

Lee

2014

2014 IEEE International Conference on Data Mining Workshop

View full text Add to dashboard Cite

Due to the ubiquity of GPS enabled devices and the advances in sensing technologies, trajectory data has become abundant. Regions of interest are important since they describe specific hot-spots within the data that often correlate with domain specific phenomena. Traditional region of interest mining utilises grid based rasters to model space. This suffers from two main problems: hard to determine the best grid size and unable to model consistent spatial adjacency. This paper utilises a 3D argument free space tessellation, Delaunay triangulation, to partition spatio-temporal trajectory data and extract arbitrary shaped regions of interest. Experimental results demonstrate the robustness and improved effectiveness of our approach at identifying granular spatio-temporal patterns.

show abstract

4D+SNN: A Spatio-Temporal Density-Based Clustering Approach with 4D Similarity

Cited by 15 publications

References 12 publications

Novel clustering and analysis techniques for mining spatio-temporal data

Novel clustering and analysis techniques for mining spatio-temporal data

A computational framework for finding interestingness hotspots in large spatio-temporal grids

Spatio-Temporal Trajectory Region-of-Interest Mining Using Delaunay Triangulation

Contact Info

Product

Resources

About