APPROXIMATE FILTERING OF PARAMETER DRIVEN PROCESSES
暂无分享,去创建一个
. A stationary stochastic process {θ,} is of interest but cannot be observed directly. Instead, at time t, we observe y, which is sampled from a distribution determined by the value of θt, say f(y∥θ,). A method for estimating the current value of {θ,} is described.
[1] Sawitzki GÜnther,et al. Finite dimensional filter systems in discrete time , 1981 .
[2] D. Gaver,et al. First-order autoregressive gamma sequences and point processes , 1980, Advances in Applied Probability.
[3] J. A. Bather,et al. Invariant Conditional Distributions , 1965 .