- [Home]
- [Research achievement]
- [Research achievement detail]
Title | On the Bayesian Forecasting Algorithm under the Non-Stationary Binomial Distribution with the Hyper Parameter Estimation (in Japanese) |
---|---|
Authors | Daiki Koizumi 、Tota Suko 、Toshiyasu Matsushima |
Released Year | 2010 |
Format | International Conference |
Category | Others |
Jounal Name | Proceeding of Ninth Valencia International Meeting on Bayesian Statistics |
Jounal Page | pp.167-168, Alicante, Spain |
Published Year | 2010 |
Published Month | 6 |
Abstract (English) |
A Bayesian forecasting algorithm under the non-stationary binomial distribution is discussed. The proposed algorithm guarantees the Bayes optimal forecasting under certain non-stationary parameter model of binomial distribution. This model can be regarded as a special case of the Simple Power Steady Model (Smith, 1979) defined as the forecasting model under the non-stationary exponential family of distribution with a known hyper parameter, but this work assumes that the hyper parameter is unknown to be estimated. Some numerical calculation results about forecasting as well as hyper parameter estimation performances would be discussed after the Bayesian forecasting method with the unknown hyper parameter is formulated. |
Note (English) |
1 |
Manuscript | |
Presentation |