An expert-supported approach to data exploration
Citation:
Cormac Hampson, 'An expert-supported approach to data exploration', [thesis], Trinity College (Dublin, Ireland). School of Computer Science & Statistics, 2011, pp. 235Download Item:
Hampson, Cormac_TCD-SCSS-PHD-2011-13.pdf (PDF) 11.09Mb
Abstract:
Almost all information domains have witnessed a large increase in the amount of structured and semi-structured data available. However, there is still a lack of support for casual computer users who wish to create queries spanning multiple information sources. Until this occurs, the real benefits of having such a proliferation of metadata will not be realised by the general public. This thesis proposes a novel expert-supported approach to data exploration that will help casual users interact with large content repositories. Specifically, this approach helps users leverage expert knowledge to discover relevant information and to draw correlations across separate data sources. These heterogeneous sources can be in various data formats, and are accessed by users in a consolidated fashion. Both a framework and a set of models based on this approach have been designed and are implemented in a technical infrastructure called SARA (Semantic Attribute Reconciliation Architecture). An associated authoring tool called SABer (Semantic Attribute Builder), which works in tandem with SARA has also been developed, and provides support for domain experts with no computing programming or data modelling experience to encode their expertise. Importantly, this means that SARA can be used in a broad range of domains, once rich metadata is available. How this expertise can then be tailored to an end-user’s interpretation or context, in order to provide him with more meaningful semantics, is another key issue tackled in this research. In summary, this thesis presents a novel and generic knowledge access platform that serves as an intermediary between curators and consumers of data. It describes the expert-supported approach to data exploration, its accompanying framework and models, as well as the implementation of SARA and SABer. Furthermore, the validation of these systems and their underlying approach is performed through five distinct evaluations. These evaluations incorporate a variety of techniques, including user trials, performance tests, questionnaires and interviews, and involve experiments with both SABer and SARA, and the third party applications that use them.
Sponsor
Grant Number
Irish Research Council for Science, Engineering and Technology. Embark
Initiative ; Science Foundation IrelandIrish Research Council for Science, Engineering and Technology. Embark
Initiative ; Science Foundation Ireland
08/IN.1/I2103
Author: Hampson, Cormac
Advisor:
Conlan, OwenQualification name:
Doctor of Philosophy (Ph.D.)Publisher:
Trinity College (Dublin, Ireland). School of Computer Science & StatisticsNote:
TARA (Trinity’s Access to Research Archive) has a robust takedown policy. Please contact us if you have any concerns: rssadmin@tcd.ieType of material:
thesisAvailability:
Full text availableLicences: