设{Xk;k≥1}是由Xk=∑i ∞=0 aiεk-i所定义的滑动平均过程,其中{εi;-∞〈i〈∞)是一同分布的φ-混合相依变量序列,{ai;i≥0)为满足条件ai-ial(i)的实数序列,l(i)为一缓变函数.当1/2〈a〈1时,{Xk;k≥1)为一长程相依过程.在Eε0 2]可能为无穷的条件下,对长程相依过程{Xk;k≥1}的部分和建立了一个更为一般性的强逼近定理.
Let {Xk;k≥1} be a moving average process defined by Xk=∑i ∞=0 aiεk-i,where {εi;-∞〈i〈∞)is a doubly infinite sequence of identically distributed φ-mixing random variables,{ai;i≥0)and l(i) is a slowly varying function. When 1/2 〈 a 〈 1, {Xk; k ≥ 1} is a long memory process. Under the assumption that Eε0 2]may be infinite, a general strong approximation theorem for partial sums of the long memory process is derived.