Bayesian Prediction of Monthly Gold Prices Using an EARSV Model and Its Competitive Component Models

The paper attempts to illustrate a methodology for the prediction of monthly gold prices using an extended autoregressive stochastic volatility (EARSV) model. The whole analysis is carried out in Bayesian paradigm by using vague priors for the parameters of the considered model. The posterior inferences for the parameters are drawn using Gibbs sampler along with intermediate Metropolis steps. The required predictions for the data set are obtained using EARSV model. Also, the predictive performance of the considered EARSV model is compared with the two competitive models. The considered EARSV model provides with more reliable results as compared to the other competitive models.