A central limit theorem with random indices for stationary linear processes (Q1802428): Difference between revisions

Consider the linear process of the form \(X_ t=\mu+\sum^ m_{j=- \infty}a_ j \varepsilon_{t-j}\), \(t \in \mathbb{Z}\), \(-\infty<\mu<\infty\), where \(\{\varepsilon_ t:t \in \mathbb{Z}\}\) is a sequence of independent and identically distributed random variables each with \({\mathcal E} \varepsilon_ t=0\), \({\mathcal E} \varepsilon^ 2_ t=\sigma^ 2<\infty\) and \(\{a_ t:t \in \mathbb{Z}\}\) is a sequence of constants satisfying \(\sum^ \infty_{j=-\infty}| a_ j |<\infty\). For such a linear process, the following central limit theorem with random indices has been proved: Let \(\{N_ n\}\) be a sequence of positive integer- valued random variables such that \(N_ n/n @>p>>N\), where \(N\) is a real- valued random variable such that \(P(0<N<\infty)=1\). Then \({\sqrt {N_ n}\{(1/N_ n) \sum^{N_ n}_{t=1}X_ t-\mu\} \over \sigma \sqrt{(\sum^ \infty_{t=-\infty} a_ t)^ 2}}\) converges in distribution to the standard normal distribution, as \(n \to \infty\). This theorem generalizes a number of well-known random central limit theorems relating to stationary linear processes.