Continuation-ratio Model for Categorical Data: A Gibbs Sampling Approach
Kai Wan · 2008
Abstract—In this paper we discuss the continuationratio model for ordinal data. This particular type of model is to model the probability of one particular category given the categories proceeding this one. It can be shown that estimation of the continuationratio model parameters can be done efficiently by using the techniques in fitting the binary data models. In this way, one does not have to estimate the cut-point parameters as in the cumulative probability models. A Bayesian approach with the use of Gibbs sampler is adopted in this paper. The adaptive rejection sampling method proposed by Gilks and Wild is used. The adaptive rejection sampling (ARS) algorithm is an efficient and direct method to sample from complicated log-concave densities often found in many Gibbs sampling scheme. We applied this model to analyse data obtained from experiments about quality of telephone connection conducted by British Telecom (BT) Laboratory. The final results are satisfactory.