Please use this identifier to cite or link to this item: http://dx.doi.org/10.18419/opus-10514
Full metadata record
DC FieldValueLanguage
dc.contributor.authorRottschäfer, Marcus Philip-
dc.date.accessioned2019-08-21T10:22:20Z-
dc.date.available2019-08-21T10:22:20Z-
dc.date.issued2019de
dc.identifier.urihttp://elib.uni-stuttgart.de/handle/11682/10531-
dc.identifier.urihttp://nbn-resolving.de/urn:nbn:de:bsz:93-opus-ds-105313de
dc.identifier.urihttp://dx.doi.org/10.18419/opus-10514-
dc.description.abstractAs more and more data becomes available through a variety of sources, for example, the Internet of Things, the need for analyzing high-dimensional data increases permanently. Visual analysis of multi-dimensional data offers an intuitive approach to better understand datasets, as humans are generally good at reasoning over spatial objects. To access a better comprehension of a dataset, data analysts often use visualization techniques in the data exploration phase. Especially convenient are dimensionality reduction techniques. They can be used for visualizations that depict the structure of the dataset and the relationship of the high-dimensional datapoints. However, visualizations based on dimensionality reduction often suffer from limited interpretability of the low-dimensional map. This is because the map does not show the relationship the low-dimensional datapoints have with the high-dimensional attributes. The Data Context Map algorithm proposed by Cheng and Mueller solves this issue by providing a low-dimensional map that accounts for the datapoint relationship, the attribute relationship, and the datapoint-attribute relationship. In this thesis, we present an implementation of the Data Context Map algorithm, enhanced by an interaction technique called Probing that allows to further reduce the error of the low-dimensional map locally. This implementation makes it possible to create visualizations of high-dimensional datasets and its attributes that improve the interpretation and the understanding of the data. Decision-making processes can be enhanced as high-dimensional data can be better analyzed. We apply the augmented Data Context Map to a variety of real-world datasets to show the advanced interpretability of the low-dimensional map.en
dc.language.isoende
dc.rightsinfo:eu-repo/semantics/openAccessde
dc.subject.ddc004de
dc.titleSimultaneous visual analysis of multidimensional data and attributesen
dc.title.alternativeSimultane Visuelle Analyse von Multidimensionalen Daten und Attributende
dc.typebachelorThesisde
ubs.fakultaetInformatik, Elektrotechnik und Informationstechnikde
ubs.institutInstitut für Visualisierung und Interaktive Systemede
ubs.publikation.noppnyesde
ubs.publikation.seiten54de
ubs.publikation.typAbschlussarbeit (Bachelor)de
Appears in Collections:05 Fakultät Informatik, Elektrotechnik und Informationstechnik

Files in This Item:
File Description SizeFormat 
2019-04-05_Camera-Ready_V1.pdf2,86 MBAdobe PDFView/Open


Items in OPUS are protected by copyright, with all rights reserved, unless otherwise indicated.