Bayesian Update for Descriptive in Fisheries Sciences


GÜNDOĞDU S. , AKAR M.

Transylvanian Review of Systematical and Ecological Research, vol.16, no.2, pp.181-192, 2013 (International Refereed University Journal)

  • Publication Type: Article / Article
  • Volume: 16 Issue: 2
  • Publication Date: 2013
  • Title of Journal : Transylvanian Review of Systematical and Ecological Research
  • Page Numbers: pp.181-192

Abstract

In this paper we examined Bayesian update for descriptive statistics for a random sample of fisheries. The Bayesian method is applied to a real sample of 730 Por’s Goat fish’s (Upeneus pori, Ben&Tuvia and Golani, 1989) length weight observation which is collected from Iskenderun Bay, Northeast Mediterranean Sea. Computational approach is to use the Markov Chain Monte Carlo simulation to draw samples from the posterior distributions of model parameters implementing the simulation in Open BUGS software. We assigned past experience as a prior distribution. This information comes from various previous studies that have been conducted in the same area. The priors for length is; theta[1]~N(11.843,1.714) and for weight is; theta[2]~N(15.815,27.321) .  According to the result, the posterior distribution for mean and variance of length were found 11.1cm and 0.003, for weight, 15.7 and 0.026. The 95% credible interval of length and weight are [10.99-11.21] and [15.42-16.05]. One of the key results of this study is that previous studies are part of the estimation and this makes variance and uncertainty low. This makes estimation sufficient and more reliable.