Statistical approximation theorems by \(k\)-positive linear operators (Q2499317)

Let \(D\) denote the space of all analytic functions on the unit disk. Then, for \(f\in D\) and \(0<r<1\), the norm in the space \(D\) is defined by \[ \|f\|_{D,r}= \max_{|z|\leq r<1}|f(z)| \] . Let, for each \(n\in N\), \(T_n:D \to D\) be a \(k\)-positive linear operator [\textit{A. D. Gadziev}, Sov. Math., Dokl. 15, 1433--1436 (1974; Zbl 0312.41013)]. Then the Taylor series of \(T_n(f)\) is given by \[ T(f;z)=\sum^\infty_{k=0}\left (\sum^\infty_{m=0}T^{(n)}_{k,m}f_m\right) z^k\qquad (n\in N), \] where \(f_m\) and, for \(k=0,1,2,\dots\) and \(n\in N\), the inner sum respectively denote Taylor coefficients of \(f\) and \(T_n(f)\). Let \(A=(a_{in})\) be a non-negative regular summability matrix and let \(x=(x_k)\). The \(A\)-statistical convergence of \(x\) to \(L\) is denoted by \(st_A-\lim x=L\) [\textit{A. R. Freedmann} and \textit{J. J. Sember}, Pac. J. Math. 95, 293--305 (1981; Zbl 0504.40002)]. In this paper, the author has proved three theorems. However, we give here his main results: Theorem. If, for each \(\nu=0,1,2\), there exists a sequence \((\varepsilon_{n,r,\nu})\) of positive real numbers such that: \(st_A-\lim_n \varepsilon_{n,r,\nu}=0\) and that \[ \left|\sum^\infty_{m=0}m^\nu T^{(n)}_{k,m}-k^\nu\right| \leq\varepsilon_{n,k,\nu},\quad n\in N,\;k=0,1,2,\dots, \] where \(T^{(n)}_{k,n}\) is the Taylor coefficient of \(T_n(f)\), then for all \(f\in D_u=\{f\in D:|f_k|\leq M_f(1+k^u)\), \(k=0,1,2,\dots,u\in Z\), the set of all integers\}, \[ st_A-\lim_n\bigl\|T_n(f)-f\bigr\|_{D,r}=0. \]