隐非齐次马尔可夫多元线性回归模型及其贝叶斯估计

Hidden Non-homogeneous Markov Multivariate Linear Regression Model and Its Bayesian Estimation

隐马尔可夫模型对于异质纵向数据的处理有良好的效果,因此被广泛应用于工程技术、生物医学、经济管理等领域。文章引入了一种特殊的非齐次隐马尔可夫状态转移方式,并将其与经典的多元线性回归相结合,提出了隐非齐次马尔可夫多元线性回归模型,介绍了对该模型进行贝叶斯推断的方法原理和技术细节。最后,通过两个模拟实验说明了推断方法的结果是可靠的。

The hidden Markov model has a good effect on processing of heterogeneous longitudinal data,so it is widely used in engineering technology,biomedicine,and economic management.This paper introduces a special non-homogeneous hidden Markov state transition mode and combines it with classical multiple linear regression to propose a hidden non-homogeneous Markov multiple linear regression model,and then introduces the method principle and technical details of Bayesian inference to this model.Finally,two simulation experiments are carried out to show that the result of the inference method is reliable.