Chiang Mai University, Thailand

Contact : Thierry Denoeux

 

In 2013 we started a collaboration with the Faculty of Economics at Chiang Mai University, Thailand, where Thierry Denoeux has been invited several times. The purpose of this collaboration is to develop and study new methods of statistical inference for estimation and prediction based on belief functions.
Although belief functions have been applied to statistics since the 1960’s, their application to complex statistical models has remained limited because of algorithmic complexity. We are developing a tractable approach based on Monte Carlo simulation. The newly developed methods are being applied to econometric models of increasing complexity, including models with a potentially large number
of covariates and temporal correlation. This research work meets some of the scientific objectives of Research topic 2.
Publications:
• O. Kanjanatarakul, S. Sriboonchitta and T. Denoeux. Forecasting using belief functions: an application to marketing econometrics. International Journal of Approximate Reasoning,
Vol. 55, Issue 5, pages 1113–1128,2014.
• K. Autchariyapanitkul, S. Chanaim, S. Sriboonchitta and T. Denoeux. Predicting stock returns in the capital asset pricing model using quantile regression and belief functions. In 3rd Int. Conf. On Belief Functions (BELIEF 2014), Oxford, UK, Springer, Sept. 2014.
• O. Kanjanatarakul, N. Kaewsompong, S. Sriboonchitta and Thierry Denoeux. Estimation and prediction using belief functions: Application to stochastic frontier analysis. In 8th conf. of the Thailand Econometric Society (TES 2015), Chiang Mai, Thailand, Springer, Jan. 2015.