Stability of the Kaczmarz Reconstruction for Stationary Sequences
Caleb Camrud,Evan Camrud,Lee Przybylski,Eric S. Weber
DOI: https://doi.org/10.1007/s11785-019-00945-8
2019-08-07
Complex Analysis and Operator Theory
Abstract:The Kaczmarz algorithm is an iterative method to reconstruct an unknown vector <em class="EmphasisTypeItalic">f</em> from inner products <span class="InlineEquation">\(\langle f , \varphi _{n} \rangle \)</span>. We consider the problem of how additive noise affects the reconstruction under the assumption that <span class="InlineEquation">\(\{ \varphi _{n} \}\)</span> form a stationary sequence. Unlike other reconstruction methods, such as frame reconstructions, the Kaczmarz reconstruction is unstable in the presence of noise. We show, however, that the reconstruction can be stabilized by relaxing the Kaczmarz algorithm; this relaxation corresponds to Abel summation when viewed as a reconstruction on the unit disc. We show, moreover, that for certain noise profiles, such as those that lie in <span class="InlineEquation">\(H^{\infty }(\mathbb {D})\)</span> or certain subspaces of <span class="InlineEquation">\(H^{2}(\mathbb {D})\)</span>, the relaxed version of the Kaczmarz algorithm can fully remove the corruption by noise in the inner products. Using the spectral representation of stationary sequences, we show that our relaxed version of the Kaczmarz algorithm also stabilizes the reconstruction of Fourier series expansions in <span class="InlineEquation">\(L^2(\mu )\)</span> when <span class="InlineEquation">\(\mu \)</span> is singular.
mathematics, applied