BAYES INFERENCE FOR TECHNOLOGICAL SUBSTITUTION DATA WITH DATA-BASED TRANSFORMATION

Bayesian inference via Gibbs sampling is studied for forecasting techn ological substitutions. The Box-Cox transformation is applied to the t ime series AR(I) data to enhance the linear model fit. We compute Baye s point and interval estimates for each of the parameters from the Gib bs sampler. The unknown parameters are the regression coefficients, th e power in the Box-Cox transformation, the serial correlation coeffici ent, and the variance of the disturbance terms. In addition, we foreca st the future technological substitution rate and its interval. Model validation and model choice issues are also addressed. Two numerical e xamples with real data sets are given.