Abstract

The Classical Binary Logistic Regression model can be used in relating Rice Mill Loss Rate and Moisture content, but the shortcomings of this model in handling cases of incomplete or sparse data will always pose a challenge. This is because, to achieve the objective of eliminating seasonal effects, data need be aggregated over each month of the year in order to develop month - specific models. This has a negative effect of reducing sample size hence posing a challenge of incomplete data. In order to resolve problems of this nature, this work employed the Bayesian Simulation Modelling Approach in relating Rice Mill Loss Rate and Moisture content for each month of the year. This was done using the Markov Chain Monte Carlo (MCMC) algorithm implemented on the Windows Bayesian Inference Using Gibbs Sampling (WINBUGS) platform. Real time data on average moisture content (%), average number of rice bags (50 kg) processed and lost in each month of the year, were sourced from MIKAP Nigeria Limited, Makurdi, Nigeria and used in the study. Major results of the study shows that optimal moisture contents of spatio-temporal characteristic will reduce Rice Mill Loss Rate from the current 5% to 1-4% and that moisture content of rough rice is a risk factor to Rice Mill Loss Rate in the months of November to May.