Help with debugging an error: Attempted to access R(1); index out of bounds because numel(R)=0???
5 views (last 30 days)
Show older comments
- Hi, I'm trying to run a bootstrap code that draws a random subsample from a dataset to use for bootstrapping. But now I get this error sequence:*
Attempted to access R(1); index out of bounds because numel(R)=0.
Error in regress (line 76)
p = sum(abs(diag(R)) > max(n,ncolX)*eps(R(1)));
Error in @(bootr)regress(Yfit+bootr,[ones(size(x,1),1),x])
Error in bootstrp (line 165)
bootstat = feval(bootfun,bootargs{:});
Error in RandomSubsampleAnalysis (line 46)
[bootstat2]=bootstrp(rep,@(bootr)regress(Yfit+bootr,[ones(size(x,1),1) x]),resid);
Caused by:
Unable to evaluate BOOTFUN with the supplied arguments.
The funny thing is that it worked fine a couple of hours ago.
my code looks like this:
clc;
warning off all;
replications=1; % number of simulations %
rep=100; % bootstrap replications %
rt1_s=[];
rt2_s=[];
OLS_b_s=[];
OLS_std_b_s=[];
Boot1_b_s=[];
Boot2_b_s=[];
Boot1_std_b_s=[];
Boot2_std_b_s=[];
for i=1:replications
% set subsample size %
ss=100; % cannot exceed full sample size %
load('data.mat');
n=length(data);
index=randsample(n,ss);
data1=data(index,:);
x=data1(:,2:end);
y=data1(:,1);
stats=regstats(y,x,'linear');
t=stats.tstat;
% case resampling
Y=[y ones(size(x,1),1) x];
[bootstat1]=bootstrp(rep,@(x)regress(x(:,1),x(:,2:end)),Y);
subplot(2,2,1), hist(bootstat1(:,1),100)
subplot(2,2,2), hist(bootstat1(:,2),100)
% model-based resampling
Yfit=[ones(size(x,1),1) x]*t.beta;
resid=y-Yfit;
[bootstat2]=bootstrp(rep,@(bootr)regress(Yfit+bootr,[ones(size(x,1),1) x]),resid);
subplot(2,2,3), hist(bootstat2(:,1),100)
subplot(2,2,4), hist(bootstat2(:,2),100)
beta = t.beta; % coefficients - OLS
se = t.se; % standard errors - OLS
Boot1_b=mean(bootstat1)'; % coefficients - case resampling
Boot1_std_b=sqrt(var(bootstat1))'; % standard errors - case resampling
rt1=abs((Boot1_b-t.beta)./Boot1_std_b);
Boot2_b=mean(bootstat2)'; % coefficients - model-based resampling
Boot2_std_b=sqrt(var(bootstat2))'; % standard errors - model-based resampling
rt2=abs((Boot2_b-t.beta)./Boot2_std_b);
rt1_s=[rt1_s;rt1'];
rt2_s=[rt2_s;rt2'];
OLS_b_s=[OLS_b_s;t.beta'];
OLS_std_b_s=[OLS_std_b_s;t.se'];
Boot1_b_s=[Boot1_b_s;Boot1_b'];
Boot2_b_s=[Boot2_b_s;Boot2_b'];
Boot1_std_b_s=[Boot1_std_b_s;Boot1_std_b'];
Boot2_std_b_s=[Boot2_std_b_s;Boot2_std_b'];
end
OLS_b=mean(OLS_b_s)'
OLS_std_b=mean(OLS_std_b_s)'
Boot1_b=mean(Boot1_b_s)'
Boot2_b=mean(Boot2_b_s)'
Boot1_std_b=mean(Boot1_std_b_s)'
Boot2_std_b=mean(Boot2_std_b_s)'
rt1=mean(rt1_s)'
rt2=mean(rt2_s)'
can someone help please?
0 Comments
Answers (1)
Stephen
on 23 Mar 2012
Are bootstrp and bootr functions you've defined? It would be really difficult to debug without seeing those. If the code used to work, and you haven't changed the functions, then I would suggest you should check the "Last Modified" date on the file 'data.mat' to see if that has changed since the last time it worked.
Good Luck,
Stephen
2 Comments
Stephen
on 26 Mar 2012
Unfortunately, I don't have the statistics toolbox so I don't have thee bootr function either. Have you tried adding breakpoints to the script and stepping in to the function to find where the issue occurs and what the value of R is when the error occurs?
Stephen
See Also
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!