Input modelling for multimodal data
Input modelling for multimodal data
Multimodal data occurs frequently in discrete-event simulation input analysis, typically arising when an input sample stream comes from different sources. A finite mixture distribution is a simple input model for representing such data, but fitting a mixture distribution is not straightforward as the problem is well-known to be statistically non-standard. Even though much studied, the most common fitting approach, Bayesian reversible jump Markov Chain Monte Carlo (RJMCMC),is not very satisfactory for use in setting up input models. We describe an alternative Bayesian approach, MAPIS, which uses maximum a posteriori estimation with importance sampling, showing it overcomes the main problems encountered with RJMCMC. We demonstrate use of a publicly-available implementation of MAPIS,which we have called FineMix, applying it to practical examples coming from finance and manufacturing.
Simulation, Mixture Models, Input modelling
1038-1052
Cheng, Russell
a4296b4e-7693-4e5f-b3d5-27b617bb9d67
Currie, Christine
dcfd0972-1b42-4fac-8a67-0258cfdeb55a
Cheng, Russell
a4296b4e-7693-4e5f-b3d5-27b617bb9d67
Currie, Christine
dcfd0972-1b42-4fac-8a67-0258cfdeb55a
Cheng, Russell and Currie, Christine
(2019)
Input modelling for multimodal data.
Journal of the Operational Research Society, 71 (6), .
(doi:10.1080/01605682.2019.1609887).
Abstract
Multimodal data occurs frequently in discrete-event simulation input analysis, typically arising when an input sample stream comes from different sources. A finite mixture distribution is a simple input model for representing such data, but fitting a mixture distribution is not straightforward as the problem is well-known to be statistically non-standard. Even though much studied, the most common fitting approach, Bayesian reversible jump Markov Chain Monte Carlo (RJMCMC),is not very satisfactory for use in setting up input models. We describe an alternative Bayesian approach, MAPIS, which uses maximum a posteriori estimation with importance sampling, showing it overcomes the main problems encountered with RJMCMC. We demonstrate use of a publicly-available implementation of MAPIS,which we have called FineMix, applying it to practical examples coming from finance and manufacturing.
Text
ChengandCurrieJORSResubmissiona
- Accepted Manuscript
More information
Accepted/In Press date: 31 March 2019
e-pub ahead of print date: 4 June 2019
Keywords:
Simulation, Mixture Models, Input modelling
Identifiers
Local EPrints ID: 430456
URI: http://eprints.soton.ac.uk/id/eprint/430456
ISSN: 0160-5682
PURE UUID: 81a6d8c8-fb93-4db3-9254-2b78dca6fbe4
Catalogue record
Date deposited: 01 May 2019 16:30
Last modified: 06 Jun 2024 04:17
Export record
Altmetrics
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
