Modelling Step Discontinuous Functions Using Bayesian Emulation

aut.embargoNoen_NZ
aut.thirdpc.containsNoen_NZ
aut.thirdpc.permissionNoen_NZ
aut.thirdpc.removedNoen_NZ
dc.contributor.advisorHankin, Robin
dc.contributor.advisorMarshall, Sarah
dc.contributor.authorAnderson, James Mark
dc.date.accessioned2017-06-13T01:47:03Z
dc.date.available2017-06-13T01:47:03Z
dc.date.copyright2017
dc.date.created2017
dc.date.issued2017
dc.date.updated2017-06-12T23:45:35Z
dc.description.abstractBayesian emulation is used in modelling complex simulators and is seen as an efficient and powerful statistical tool. The simulations can be very time-consuming to run, resulting in only being able to run the simulator a limited number of times. From the literature, it is has been suggested that the emulator is best suited for continuous functions; however, it is very common to find physical problems containing discontinuities. These discontinuities' positions may also be unknown and therefore, for this research, information on where the discontinuities will be withheld from the emulator. This thesis focuses on emulating the Heaviside function as one simple function containing step-discontinuities and then progresses to slightly more complex functions with step-discontinuities. Specific goodness-of-fit measure have been designed to highlight and measure the emulator when applied to theses step-discontinuous, such as mean squared error and a simple design of Jaccard index. The numerical calculations of the goodness-of-fit techniques are carried out in the R statistical programming language, with the BACCO package for the emulator. It is found that the emulator is able to model discontinuous functions to some degree but, unless there is certainty about the discontinuities' locations, care should be taken when using Bayesian emulation to model discontinuous functions.en_NZ
dc.identifier.urihttps://hdl.handle.net/10292/10543
dc.language.isoenen_NZ
dc.publisherAuckland University of Technology
dc.rights.accessrightsOpenAccess
dc.subjectGaussian process emulatoren_NZ
dc.subjectBayesianen_NZ
dc.subjectEmulationen_NZ
dc.subjectEmulatoren_NZ
dc.subjectStep Discontinuousen_NZ
dc.subjectHeavisdieen_NZ
dc.subjectSimulationen_NZ
dc.subjectSimulatoren_NZ
dc.subjectGaussian processen_NZ
dc.subjectGoodness-of-fiten_NZ
dc.subjectBACCOen_NZ
dc.titleModelling Step Discontinuous Functions Using Bayesian Emulationen_NZ
dc.typeThesis
thesis.degree.grantorAuckland University of Technology
thesis.degree.levelMasters Theses
thesis.degree.nameMaster of Scienceen_NZ
Files
Original bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
AndersonJ1.pdf
Size:
4.07 MB
Format:
Adobe Portable Document Format
Description:
Thesis
License bundle
Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
license.txt
Size:
897 B
Format:
Item-specific license agreed upon to submission
Description:
Collections