A bootstrap method for estimating bias and variance in statistical fisheries modelling frameworks using highly disparate datasets

B. Th Elvarsson, L. Taylor, V. M. Trenkel, V. Kupca, G. Stefansson

Research output: Contribution to journalArticlepeer-review

Abstract

Statistical models of marine ecosystems use a variety of data sources to estimate parameters using composite or weighted likelihood functions with associated weighting issues and questions on how to obtain variance estimates. Regardless of the method used to obtain point estimates, a method is required for variance estimation. A bootstrap technique is introduced for the evaluation of uncertainty in such models, taking into account inherent spatial and temporal correlations in the datasets, which are commonly transferred as assumptions from a likelihood estimation procedure into Hessian-based variance estimation procedures. The technique is demonstrated on a real dataset and the effects of the number of bootstrap samples on estimation bias and variance estimates are studied. Although the modelling framework and bootstrap method can be applied to multispecies and multiarea models, for clarity the case study described is of a single-species and single-area model. © 2014

Original languageEnglish
Pages (from-to)99-110
Number of pages12
JournalAfrican Journal of Marine Science
Volume36
Issue number1
DOIs
Publication statusPublished - Jan 2014

Bibliographical note

Funding Information: undertaken while authors BÞE, LT, VK and GS were employed at the MRI (Marine Research Institute, Reykjavik) and uses data from the MRI databases. The Gadget code has been in development for more than a decade by many programmers at the MRI and IMR (Institute of Marine Research, Bergen). The work was supported in part by EU grants QLK5-CT1999-01609 and FP6 TP8.1 502482, as well as a grant from The Icelandic Centre for Research (Rannis). The authors would like to thank Dr S Gavaris for useful discussions, which have considerably improved the paper.

Other keywords

  • bootstrapping
  • correlated data
  • fish population dynamics
  • non-linear models

Fingerprint

Dive into the research topics of 'A bootstrap method for estimating bias and variance in statistical fisheries modelling frameworks using highly disparate datasets'. Together they form a unique fingerprint.

Cite this