Abstract:
General Rogosinsky–Bernstein linear polynomial means $R_n(f)$ of Fourier series are introduced and three convergence criteria as $n\to\infty$ are obtained: for convergence in the space $C$ of continuous periodic functions and for convergence almost everywhere with two guaranteed sets (Lebesgue points and $d$-points). For smooth functions, the rate of convergence in norm of $R_n(f)$, as well as of their interpolation analogues, is also studied. For approximation of functions in $C^r$, the asymptotics is found along with the rate of decrease of the remainder term.
Keywords:series and Fourier transforms, Hardy's inequality, Riesz means, Lebesgue points ($l$-points) and $d$-points, modulus of smoothness, linearized modulus of smoothness, Jackson's theorem, Vallée-Poussin polynomial, conjugate function, entire functions of exponential type, comparison principle, Marcinkiewicz's inequality and discretization.