Application of Bayesian fuzzy regression analysis and model selection on Hedonic Pricing Strategy / Mohamad Idham Md Razak … [et al.]

• Main Authors: Md Razak, Mohamad Idham, Omar, Roaimah, Mohammad Amin, Nor Azizah, Norhisham, Norshiba
• Format: Book Section
• Language: English
• Published: Division of Research, Industrial Linkages and Alumni, UiTM Cawangan Melaka 2013
• Subjects: Integer programming
• Online Access: https://ir.uitm.edu.my/id/eprint/80512/1/80512.pdf; https://ir.uitm.edu.my/id/eprint/80512/
• Institution: Universiti Teknologi Mara

In the following prototype it has been suggested a Bayesian approach to fuzzy clustering analysis - the Bayesian fuzzy regression. Bayesian Posterior Odds analysis is employed to select the correct number of clusters for the fuzzy regression analysis. In this study, we use a natural conjugate prior for the parameters, and we find that the Bayesian Posterior Odds provide a very powerful tool for choosing the number of clusters. The results from a Hedonic Pricing Strategy experiment and two real data applications of Bayesian fuzzy regression are very encouraging. Recent developments in econometric modelling have emphasized a variety of non-linear specifications. Parametric examples of these include various regime-switching models (the threshold and Markov switching autoregressive models): threshold autoregressive (TAR) models, self-exciting threshold autoregressive (SETAR) models, smoothing threshold autoregressive (STAR) models, and various others. In addition, non-parametric and semi-parametric models are widely used, though the well-known "curse of dimensionality" can place some limitations on their use with multivariate data. The use of fuzzy clustering analysis in the context of econometric modelling is a rather new approach within the class of nonlinear econometric models.