Code covered by the BSD License  

Highlights from
Sparse Randomized Kaczmarz for Multiple Measurement Vectors.

Sparse Randomized Kaczmarz for Multiple Measurement Vectors.

by

 

Implementation of sparse randomized Kaczmarz algorithm to handle multiple measurements.

Demo.m
%This is a simple demo of our implementation of Sparse Randomized Kaczmarz
%algorithm to handle multiple measurement vectors.
%please see the function for commented details on input and output.
%No. of multiple measurement vectors are represented by L
% A is m by n
% X is n by L   
% B = AX  is m by L 
clc; clear all ; close all;

m=500;  n=100; L=5; K=10; estSupp=20; J=5; %overdetermined system

    AvgOver=1;   %Repeats the experiment this many time
    A=NormalMatrix(m,n);
    Xorg=zeros(n,L);
    temp=randperm(n);
    Xorg(temp(1:K),:)=randn(K,L);
    B=A*Xorg;

    errSrk=zeros(AvgOver,1);
    

    for i=1:AvgOver         
        Xrec=srkMMV(A,B,J,estSupp,'last');
        errSrk(i)= norm(Xorg-Xrec,'fro')/norm(Xorg,'fro');           

    end
        
   fprintf('\n Mean Error with SRK-MMV = %1.8f \n',mean(errSrk));
   

Contact us