University of Twente Student Theses
Clustering spatio-temporal movement data
Sarhangi, Amir (2010) Clustering spatio-temporal movement data.
PDF
3MB |
Abstract: | Geographic Information (GI)community faced to the increase in availability of data during last decades due to advancement in the collection devices and also user requirements. Processing and extracting meaningful information from this large amount of data is the domain of KDD and currently Geovisual Analytics. Clustering in all these concepts plays an important role. Spatio temporal movement data is one kind of data, in which interest is growing too fast. Clustering of spatio temporal movement data is quite a new area of research, which extends the existing clustering algorithms to appropriate use. OPTICS is a sophisticated algorithm which has been used mainly in trajectory clustering while the focuss of this research is mainly on point based clustering by the aid of 3D Euclidian distance function. In this work, original codes of OPTICS were extended to work in space-time, and implemented using the JAVA programming language. The algorithm tested over three human movement data sets: GPS, RFID and cell phone tracking data. The results are evaluated by visual interactive techniques and validity indices. In all data sets spatio-temporal clusters are achieved. Keywords Spatio-temporal data, clustering, Human movement, Geovisual analytics |
Item Type: | Essay (Master) |
Faculty: | ITC: Faculty of Geo-information Science and Earth Observation |
Programme: | Geoinformation Science and Earth Observation MSc (75014) |
Link to this item: | https://purl.utwente.nl/essays/92482 |
Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Repository Staff Only: item control page