2014년 제1회 통계세미나 개최 안내

- BK 세미나 시리즈 -

 

통계연구소에서는 다음과 같이 통계 세미나를 개최하오니 많은 참여 바랍니다.

 

일시: 2014225() 오후 5

장소: 고려대학교 정경관 508

연사: Professor Fumiyasu Komaki (University of Tokyo)

 

Prior construction for Bayesian prediction

 

Bayesian predictive densities for future observables are considered, and construction methods for prior densities are discussed from a predictive viewpoint. The performance of predictive densities is evaluated by the Kullback-Leibler divergence. Asymptotic properties of Bayesian predictive densities are investigated from the viewpoint of information geometry. Construction methods of objective priors by using conditional mutual information between future observables and unknown parameters given observations are also discussed, if time permits.

 

고려대학교 통계연구소