How can I define a foor loop in which there is a bootci function?
Show older comments
Hello. Firstly I estimate cubic splines coefficients and then I would like to implement bootstrap function on each interval of cubic splines. This is my code and I got the error " Not enough input arguments ".
t=0:.1:pi;
y=sin(t);
N=length(t); %number of points
n=N-1; %number of subintervals
h=(t(N)-t(1))/n; %step size
% h=2;
Trid=diag(4*ones(1,n-1))+diag(ones(1,n-2),1)+diag(ones(1,n-2),-1);
for i=1:n-1
z(i)=6/h^2*(y(i+2)-2*y(i+1)+y(i));
end
z=z';
w=inv(Trid)*z;
sigma=[0;w;0];
for i=1:n
d(i)=y(i);
b(i)=sigma(i)/2;
a(i)=(sigma(i+1)-sigma(i))/(6*h);
c(i)=(y(i+1)-y(i))/h-h/6*(2*sigma(i)+sigma(i+1));
end
r=1; %number of subintervals
hh=h/r; %step size of subintervals
x=t(1):hh:t(N);
for i=1:n
for j=r*(i-1)+1:r*i
s(j)=a(i)*(x(j)-t(i))^3+b(i)*(x(j)-t(i))^2+c(i)*(x(j)-t(i))+d(i);
end
end
s(r*n+1)=y(N);
for i=1:n
s_i{i}=@(x) a(i)*(x-t(i)).^3+b(i)*(x-t(i)).^2+c(i)*(x-t(i))+d(i);
end
k=1;
for i=1:n
B = [-.161 0 1 0];
beta0(i,:)=[a(i) b(i) c(i) d(i)];
x=t';
s_i{i}=@(x) a(i)*(x-t(i)).^3+b(i)*(x-t(i)).^2+c(i)*(x-t(i))+d(i);
modelfunc(i)=s_i{i};
beta(i) = @(predictor,response)nlinfit(predictor,response,modelfunc(i),beta0(i,:))
c(:,k) = bootci(1000,beta(i),t',y,'Type','norm');
k=k+2;
end
Accepted Answer
More Answers (0)
Categories
Find more on Resampling Techniques in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!