A strong invariance principle concerning the \(J\)-upper order statistics for stationary Gaussian sequences (Q2365738)

Consider a centred stationary Gaussian sequence \(\{X_ n,- \infty<n<\infty\}\); let \(X_{1,n}<X_{2,n}<\cdots<X_{n,n}\) be the order statistics of \(\{X_ 1,\dots,X_ n\}\). For a given integer \(J\geq 1\), let \(\{T_ k, k\geq 1\}\) be the \(J\)-th record times and let \(\Theta_ k=\{X_{T_ k+1-i}: 1\leq i\leq T_ k\}\). Then \(\{T_ k,\Theta_ k\}\) is called the \(J\)-th record sequence based on \(\{X_ n\}\). The main result presents three hypotheses, any one of which guarantees the existence of a probability space on which both \(\{X_ n\}\) and a sequence of independent standard normal random variables \(\{X_ n^*, n\geq 1\}\) are defined such that almost surely integers \(N\) and \(q\) exist such that \(S_ n=T_{n-q}\) and \(R_ n=\Theta_{n-q}\) for every \(n\geq N\), where \((S_ k,R_ n)\) is the \(J\)-th record sequence based on \(\{X_ n^*\}\). This result generalizes some work of \textit{G. Haiman} [Mathematical statistics and probability theory, Vol. A, Proc. 6th Pannonian Symp., Bad Tatzmannsdorf/Austria 1986, 105-120 (1987; Zbl 0654.60030)].