Abstract
Modeling and forecasting ordinal outcomes has become a core study for many statisticians because of the many forms of data encountered in real life, that have such a format. Several authors have proposed different approaches in modeling this type of data either in classical approaches (Mc Cullagh, 1980) or from a Bayesian perspective (Albert and Chib, 1993) (Cowles et al., 1996). One commonly adopted method of modeling ordinal data is that the observed ordinal scores have a correspondence with the latent variable through a set of cut points. Sometimes there are some difficulties in estimation. One of these categories is koto. (Albert and Chib, 1993) proposed an ordinal model in a Bayesian framework that fuzzy-collaborates prior on the cut-point parameters. The approach is used to estimate these parameters through their posterior distribution. We then compare our results after prediction with ordinal logistic regression to see which methods through which we obtain the best estimates.
