Please use this identifier to cite or link to this item:
http://theses.ncl.ac.uk/jspui/handle/10443/664
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Williams, Michael Kevin | - |
dc.date.accessioned | 2010-03-12T09:16:02Z | - |
dc.date.available | 2010-03-12T09:16:02Z | - |
dc.date.issued | 1995 | - |
dc.identifier.uri | http://hdl.handle.net/10443/664 | - |
dc.description | PhD Thesis | en_US |
dc.description.abstract | Statistical expert systems have grown more numerous over the last few years. However, the tools which are used in their development are largely modifications of existing methods from other areas. This means that the unique aspects of statistical problems may be omitted in their development. This thesis examines the way that a statistical consultant works, and proposes guidelines and tools for system development based on this. Section I provides a brief introduction to this area. Section II considers the problems faced in establishing the actual problem brought by a client. Two areas of particular interest are establishing the background subject and the facts of a problem. Two methods are proposed to tackle these areas. Also of interest is the way in which a consultant may select a statistical test. Section III introduces a program to elicit the rules used by a consultant. A disk containing the program and the functions used is included with this thesis. | en_US |
dc.description.sponsorship | Science and Engineering Research Council | en_US |
dc.language.iso | en | en_US |
dc.publisher | Newcastle University | en_US |
dc.title | The construction of statistical expert systems with applications | en_US |
dc.type | Thesis | en_US |
Appears in Collections: | School of Mathematics and Statistics |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Williams 95.pdf | Thesis | 7.6 MB | Adobe PDF | View/Open |
dspacelicence.pdf | Licence | 43.82 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.