摘要
给出了观测值服从多项式分布的动态多项式分布模型, 并在自然参数Bt 与状态参数θt之间满足f(Bt) = F′tθt 的假设下,
In this paper the dynamic polynomial distribution model for observations following polynomial distribution is proposed. Assuming the relation \$f(B\-t)=F\+′\-tθ\-t\$ between the natural parameter and system parameter, the updating recursion of the model parameters is given on the basis of conjugate distribution. And Bayesian forecasting is derived, too.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
1999年第10期122-124,共3页
Systems Engineering-Theory & Practice
基金
山东省自然科学基金
关键词
多项式
分布模型
BAYES预测
dynamic polynomial distribution model
Bayesian forecasting