Linear Prediction Theory and Markov Models of a Homogeneous Random Field with Discrete Parameters (II)
Generally, the definations of the Markov type for homogeneous random field as follows: Let $H_X(T)$ be the closed linear manifold spanned by all $X(m,n), (m,n)\in T, T_0\subset T$. $(m',n')\bar{\in}T$, if $P_{H_X(T)}X(m',n')=P_{H_X(T_0)}X(m',n')$, then we say that $H_X(T)$ has the Morkov property for $H_X(T_0)$ at $X(m',n')$. In this paper, three types are posed and discussed: 1). \$T = \{(m,n), -\infty