Adjusting beliefs via transformed fuzzy prices

Abstract

© 2018, Springer International Publishing AG. In the situation that the gut feeling tells otherwise, incorporating information from expert opinions can significantly improve the accuracy of standard estimation and prediction methods, which rely only on observed data. To cope with this problem, we propose the fusion of data under the Bayesian framework by transforming price estimates into initial beliefs of assets. The proposed methodology focuses on modeling the price expectation by linguistic terms and mathematically extending them to other parameters like the standard deviation. On five sample assets from different markets, our method was experimented and compared with the method of the ARMA-GARCH beyond the points of structural change. The problems are multi-dimensional but conveniently solved by the Metropolis-Hastings algorithm. The results show the significant impacts of expert opinions on the posterior.