Integrating information about complex systems: role of metadata in the acceptability of results from models.

Westlake, Andrew and Wigan, Marcus (2007) Integrating information about complex systems: role of metadata in the acceptability of results from models. Transportation Research Record, 1972 (2006). pp. 60-68. ISSN 0361-1981

Full text not available from this repository. (Request a copy)


Opus is a project funded by the European Union Statistical Agency (Eurostat). The project is developing a methodology for the coherent and consistent integration of information from multiple sources about complex systems (based on Bayesian statistical models) and a series demonstration of this approach. Most, but not all, of these are transport applications. Alongside the modeling frameworks being developed, the project has developed a framework for storing a complete audit trail, covering the specification and fitting of statistical models, and making use of a formal metadata specification to represent statistical models. This framework builds on ideas about metadata for processes that have been developed in other statistical projects. Practitioners are often skeptical about results derived from models. In practice, all analysis of survey or sample data involves some form of model, even if it is hidden in implicit assumptions. The approach to this skepticism is to open up the model to scrutiny and to present information about the reliability of results, under the title of provenance and reliability. All this comes from the metadata about the statistical model. The approach is generic, but the presentation needs to be domain specific and tailored to the level of expertise and understanding of the user of the results. This paper presents the motivation behind this approach and outlines the metadata structures and functionality being built to deliver this support to users of results from the Opus methodology.

Item Type: Article
Print ISSN: 0361-1981
Uncontrolled Keywords: Opus; Eurostat; Information management; Transport applications; Data integration; Bayesian statistical models; Data provenance audit trails;
University Divisions/Research Centres: Edinburgh Napier University, Transport Research Institute
Dewey Decimal Subjects: 500 Science > 510 Mathematics > 510 Mathematics
000 Computer science, information & general works > 000 Computer science, knowledge & systems > 004 Data processing & computer science
300 Social sciences > 380 Commerce, communications & transportation > 388 Transportation; ground transportation
Library of Congress Subjects: H Social Sciences > HE Transportation and Communications
H Social Sciences > HA Statistics
Q Science > QA Mathematics > QA76 Computer software
Item ID: 1933
Depositing User: RAE Import
Date Deposited: 22 Feb 2008 14:54
Last Modified: 11 Dec 2015 10:51

Actions (login required)

View Item View Item

Edinburgh Napier University is a registered Scottish charity. Registration number SC018373