Inference in Complex Systems Using Multi-Phase MCMC Sampling With Gradient Matching Burn-in (original) (raw)
Lazarus, Alan, Husmeier, Dirk ORCID: https://orcid.org/0000-0003-1673-7413 and Papamarkou, Theodore
ORCID: https://orcid.org/0000-0002-9689-543X(2017) Inference in Complex Systems Using Multi-Phase MCMC Sampling With Gradient Matching Burn-in. In: 32nd International Workshop on Statistical Modelling, Groningen, Netherlands, 03-07 Jul 2017, pp. 52-57.
| [ |
Text 144287.pdf - Published Version 3MB |
|---|
Publisher's URL: https://iwsm2017.webhosting.rug.nl/IWSM_2017_V1.pdf
Abstract
We propose a novel method for parameter inference that builds on the current research in gradient matching surrogate likelihood spaces. Adopting a three phase technique, we demonstrate that it is possible to obtain parameter estimates of limited bias whilst still adopting the paradigm of the computationally cheap surrogate approximation.
| Item Type: | Conference Proceedings |
|---|---|
| Status: | Published |
| Refereed: | Yes |
| Glasgow Author(s) Enlighten ID: | Papamarkou, Dr Theodore and Lazarus, Mr Alan and Husmeier, Professor Dirk |
| Authors: | Lazarus, A., Husmeier, D., and Papamarkou, T. |
| College/School: | College of Science and Engineering > School of Mathematics and Statistics > Statistics |
| Copyright Holders: | Copyright © 2017 The Authors |
| First Published: | First published in Proceedings of the 32nd International Workshop on Statistical Modelling: 52-57 |
| Publisher Policy: | Reproduced with the permission of the Editor |
University Staff: Request a correction | Enlighten Editors: Update this record
Funder and Project Information
1
Computational inference in systems biology
Dirk Husmeier
EP/L020319/1
M&S - STATISTICS
Deposit and Record Details
| ID Code: | 144287 |
|---|---|
| Depositing User: | Professor Dirk Husmeier |
| Datestamp: | 18 Jul 2017 08:19 |
| Last Modified: | 02 May 2025 14:41 |
| Date of first online publication: | 3 July 2017 |
| Date Deposited: | 18 July 2017 |
| Data Availability Statement: | No |