Convergence analysis of a class of adaptive weighted norm extrapolation algorithms

Abstract

Adaptive weighted norm extrapolation algorithms can provide superior performance for estimation of sparse signals from limited data. We present theoretical analysis results for a class of these algorithms that include a proof of the global convergence, the rate of convergence derivation, and characterization of the fixed points. We also propose a general class of adaptive weighted extrapolation algorithms and introduce a more general problem formulation which greatly expands the range of applications of the algorithm.<>