非线性Bayes动态模型的研究

A RESOLUTION TO NONLINEAR BAYESIAN DYNAMIC MODELS

对正态误差下的非线性Bayes动态模型进行了处理 .对模型分观测误差已知和未知常数两种情况进行了研究 ,利用Taylor展开来近似参数分布 ,并引入E -M算法 ,完成了模型的修正。

Nonlinear bayesian dynamic models under normal errors are resolved.We study two kinds of models whose observation variances are known and unknown constant.The distribution of model parameters is obtained through Taylor expansion,and the updating and recursion of model are completed using E-M algorithm.