jackknife

 Compute jackknife estimates of the mean and standard deviation of input data x
 Usage: [m,jsd]=jackknife(x)
 
 Inputs:
 x : data in the form samples x trials

 Outputs:
 m : estimate of the mean (across trials)
 jsd: jackknife estimate of the standard deviation (across trials)

0001 function [m,jsd]=jackknife(x)
0002 % Compute jackknife estimates of the mean and standard deviation of input data x
0003 % Usage: [m,jsd]=jackknife(x)
0004 %
0005 % Inputs:
0006 % x : data in the form samples x trials
0007 %
0008 % Outputs:
0009 % m : estimate of the mean (across trials)
0010 % jsd: jackknife estimate of the standard deviation (across trials)
0011 
0012 [N,C]=size(x);
0013 if C==1; error('Need multiple trials'); end;
0014 m=mean(x,2);
0015 theta=zeros(N,C);
0016 for tr=1:C;
0017     i=setdiff((1:C),tr); % drop 1 trial
0018     y=sum(x(:,i),2)/(C-1); % mean over remaining trials
0019     theta(:,tr)=C*m-(C-1)*y; % pseudo values
0020 %     yy(:,tr)=y;
0021 end;
0022 jm=mean(theta,2);
0023 jm=repmat(jm,[1 C]);
0024 % jm2=mean(yy,2);
0025 % jm2=repmat(jm2,[1 C]);
0026 jsd=sqrt(sum((theta-jm).^2,2)/(C*(C-1)));
0027 % jsd2=sqrt((C-1)*sum((yy-jm2).^2,2)/C);
0028 % jsd
0029 % jsd2