Behavior of the Fourier-Walsh coefficients of a corrected function (Q311254)

Suppose that \(n\in\mathbb Z_+=\{0,1,\dots\}\) has a binary expansion \(n=\sum^\infty_{k=0}\nu_k2^k\), \(\nu_k\in \{0,1\}\). Every \(x\in [0,1)\) may be represented in the form \(x=\sum^\infty_{k=0}x_k/2^{k+1}\), \(x_k\in\{0,1\}\) (for \(x=l/2^n\), \(n\in\mathbb N\), \(l\in (0,2^n)\cap \mathbb Z\), we consider the expansion with finite number of \(x_k\neq 0\)). Then the \(n\)-th Walsh function is defined by \[ w_n(x)=(-1)^{\sum^\infty_{k=0}\nu_kx_k}, \quad n\in\mathbb Z_+. \] For \(f\in L^1[0,1)\) we consider the Walsh-Fourier coefficients of \(f\) by \(c_k(f)=\int^1_0f(x)w_k(x)\,dx\), \(k\in\mathbb Z_+=\{0,1,\dots\}\), and \(\mathrm{spec}(f)=\{k\in\mathbb Z_+: c_k(f)\neq 0\}\). The main result of the paper is Theorem. Consider a sequence \(\{a_k\}^\infty_{k=1}\) with \(a_k\downarrow 0\) and \(\{a_k\}^\infty_{k=1}\notin l^2\). If \(0<\delta<1\), \(p\in[1,2]\) and \(f\in L^p[0,1)\) are given, then there exists \(\widetilde{f}\in L^p[0,1)\) such that \(f(x)\neq \widetilde{f}(x)\) on a set of measure less than \(\delta\) and the non-zero Walsh-Fourier coefficients of \(\widetilde{f}\) satisfy the equality \(|c_k(\widetilde{f})|=a_k\), \(k\in \mathrm{spec}(\widetilde{f})\).