Statistical limit point theorems (Q1578977)

A sequence \(x\) of complex numbers is said to be statistically convergent to \(L\) if for every \(\varepsilon >0\), the natural density of the set \(\{k: |x_k - L|\geq \varepsilon \}\) is \(0\). It is known that \(x\) is statistically convergent if and only if there exists a convergent sequence \(y\) such that \(x_k=y_k\) for almost all \(k\), i.e. on a set with natural density \(1\). Furthermore, a number \(\lambda\) is said to be a statistical limit point of \(x\) if there exists a subsequence \(z=(x)_K\) of \(x\) such that \(K\subseteq \mathbb{N}\) does not have natural density \(0\) and \(z\) converges to \(\lambda\). The author gives statistical convergence analogues of the following results, cf. \textit{R. P. Agnew} [Bull. Am. Math. Soc. 50, 596-598 (1944; Zbl 0060.16002)], \textit{J. A. Fridy} [Math. Z. 143, 187-192 (1975; Zbl 0292.40002)] and \textit{D. F. Dawson} [Pac. J. Math. 44, 455-460 (1973; Zbl 0256.40004]: If \(A\) is a regular matrix and \(x\) a bounded sequence, then there is a subsequence (rearrangement, stretching, respectively) \(z\) of \(x\) such that the set of limit points of \(Az\) includes the set of limit points of \(x\).