Advanced

Modelling of African Farm Dynamics Using Bivariate Binary Logistic Regression in WinBUGS

Ali, Ghazanfar and Darda, Md Abud (2009)
Department of Statistics
Abstract
The recent development and implementation of Markov Chain Monte Carlo (MCMC) method is generating application of complicated models over a wide range of sciences. For two correlated binary responses, bivariate binary logistic regression is a suitable way to identify the related covariates and at
the same time, their interactions validity can be investigated in terms of logarithm of odds ratio. Present study uses data obtained from a longitudinal survey conducted in 2002-2005 to make a guideline for the agricultural development and food security in Africa. Defining the term Extensification and Intensification as two traditions for the farm dynamics in the selected African
states, a simple data analysis code for bivariate binary regression... (More)
The recent development and implementation of Markov Chain Monte Carlo (MCMC) method is generating application of complicated models over a wide range of sciences. For two correlated binary responses, bivariate binary logistic regression is a suitable way to identify the related covariates and at
the same time, their interactions validity can be investigated in terms of logarithm of odds ratio. Present study uses data obtained from a longitudinal survey conducted in 2002-2005 to make a guideline for the agricultural development and food security in Africa. Defining the term Extensification and Intensification as two traditions for the farm dynamics in the selected African
states, a simple data analysis code for bivariate binary regression analysis in WinBUGS is developed and comparison is made with the analysis obtained in R under maximum likelihood estimation. Results indicate that some factors for instance ‘availability of new crop technology’, ‘import of Maize’ and ‘stopped intercropping’ shows some negative association with farm dynamics response variable, which concludes that these factors discouraging the production of Maize and areal increase in both Bayesian and maximum likelihood estimation approach. Whereas ‘Change in fertilizer use’, ‘cultivated area increase’ and ‘started selling maize’ shows positive association. This indicates that these factors support the argument of areal and Maize production increase. Farm holders access to modern crop technologies, in combination with commercial incentives to staple crop production emerge as the most important explanation of dynamism. Thus for independent modelling of Extensification and Intensification dynamics, both Bayesian and frequentistic approach mimics the result. But the joint association provides distinctive result in Bayesian approach that concludes that Extensification and Intensification are two diverse way of farm dynamics. (Less)
Please use this url to cite or link to this publication:
author
Ali, Ghazanfar and Darda, Md Abud
supervisor
organization
year
type
H1 - Master's Degree (One Year)
subject
keywords
WinBUGS, Bivariate binary logistic regression, Extensification, Intensification, Statistics, operations research, programming, actuarial mathematics, Statistik, operationsanalys, programmering, aktuariematematik
language
Swedish
id
1848808
date added to LUP
2009-10-22 00:00:00
date last changed
2011-06-01 12:43:57
@misc{1848808,
  abstract     = {The recent development and implementation of Markov Chain Monte Carlo (MCMC) method is generating application of complicated models over a wide range of sciences. For two correlated binary responses, bivariate binary logistic regression is a suitable way to identify the related covariates and at
the same time, their interactions validity can be investigated in terms of logarithm of odds ratio. Present study uses data obtained from a longitudinal survey conducted in 2002-2005 to make a guideline for the agricultural development and food security in Africa. Defining the term Extensification and Intensification as two traditions for the farm dynamics in the selected African
states, a simple data analysis code for bivariate binary regression analysis in WinBUGS is developed and comparison is made with the analysis obtained in R under maximum likelihood estimation. Results indicate that some factors for instance ‘availability of new crop technology’, ‘import of Maize’ and ‘stopped intercropping’ shows some negative association with farm dynamics response variable, which concludes that these factors discouraging the production of Maize and areal increase in both Bayesian and maximum likelihood estimation approach. Whereas ‘Change in fertilizer use’, ‘cultivated area increase’ and ‘started selling maize’ shows positive association. This indicates that these factors support the argument of areal and Maize production increase. Farm holders access to modern crop technologies, in combination with commercial incentives to staple crop production emerge as the most important explanation of dynamism. Thus for independent modelling of Extensification and Intensification dynamics, both Bayesian and frequentistic approach mimics the result. But the joint association provides distinctive result in Bayesian approach that concludes that Extensification and Intensification are two diverse way of farm dynamics.},
  author       = {Ali, Ghazanfar and Darda, Md Abud},
  keyword      = {WinBUGS,Bivariate binary logistic regression,Extensification,Intensification,Statistics, operations research, programming, actuarial mathematics,Statistik, operationsanalys, programmering, aktuariematematik},
  language     = {swe},
  note         = {Student Paper},
  title        = {Modelling of African Farm Dynamics Using Bivariate Binary Logistic Regression in WinBUGS},
  year         = {2009},
}