Abstract Factor analysis model plays an important role in characterizing dependence of latent factors on the observed variables and interpreting the correlation among the observed indexes (variables). However, in real applications, data set often takes on the temporal variability, multimode, skewness, and so on. In this paper, we extended the classic factor analysis model to the dynamic factor model mixed with homogenous hidden Markov model, and developed a Bayesian semiparametric analysis procedure. Blocked Gibbs sampler is used to implement posterior sampling. The empirical results show that our method is effective.
XIA Ye-mao, GOU Jian-wei, LIU Ying-an. Semi-parametric Bayesian analysis for factor analysis model mixed with hidden Markov model. Applied Mathematics A Journal of Chinese Universities, 2015, 30(1): 17-30.
