BAYESIAN APPROACH TO MODELLING THE RELATIONSHIP BETWEEN ORANGE ROT RATE AND TRANSPORTATION DISTANCE

Abstract

A Bayesian Logistic Regression Model relating the rot rate of oranges and transportation distance using incomplete data was developed. The incompleteness of data limits the use of the
Classical Logistic Regression Model and this limitation informed our development of a Bayesian Statistical Simulation Modelling Procedure. This work used the simulation procedure through the Markov Chain Monte Carlo (MCMC) algorithm executed in WINBUGS. It is found that transportation distance has a linear and positive relationship with orange rot rate. For both rainy and dry seasons, a kilometre increase in transportation distance will increase the odds that oranges transported from Benue State to Gombe Fruit Market will be rotten. It was also discovered that the mean rot rate of oranges transported from Benue State during the rainy season is 19% and 13% during the dry season. The researcher noted that the distribution of the estimated unknown parameter ( )  converges to the posterior distribution quickly indicating model adequacy. The model also predicted the best distance for maximizing profitability and minimising losses.