The central limit theorem for \(\rho\)-mixing series patterns (Q1916577)

The article gives a version for \(\rho\)-mixing triangular arrays \(\{\xi_{nk}: k= 1,\dots, k_n, n\geq 1\}\), \(k_n\to \infty\) as \(n\to \infty\), of the classical sufficient conditions for convergence to the normal law under the assumption that \(\{\xi_{nk}\}\) satisfies \[ \rho(1)< 1\quad \text{and} \quad \sum^\infty_{n= 1} \rho(2^n)< \infty, \tag{\(*\)} \] \(\rho(k)\) being the maximal correlation coefficient of \(\{\xi_{nk}\}\). The theorem obtained generalizes the results on the convergence of the normalized partial sums of stationary sequences satisfying \((*)\) which have been proved by \textit{I. A. Ibragimov} [Theory Probab. Appl. 20, 135-141 (1975); translation from Teor. Veroyatn. Primen. 20, 134-140 (1975; Zbl 0335.60023)] and by \textit{R. C. Bradley} [Ann. Probab. 16, No. 1, 313-332 (1988; Zbl 0643.60018)], in the finite and infinite variance case, respectively.