How to compute the standard error for coefficients returned from fit function?
2 views (last 30 days)
Show older comments
I fit my data to an exponential model using the FIT function:
% X data
for i=1:32
TE(i)=i*12.4e-3;
end
X=TE(2:end);
% Y data
MeanA=[1.2022e+03, 1.1752e+03, 969.7900, 851.6100, 708.6800, 621.5000, 515.5300, 454.1000, 362.2100, 334.1400, 269.0400, 236.8100, 216.3000, 201.4900, 163.6800, 154.6000, 156.5600, 140.9800, 143.9400, 157.7100, 115.1300, 114.9200, 115.6300, 111.0800, 104.1600, 105.5900, 110.0500, 112.1600, 127.5700, 107.1800, 111.8200, 95.7400];
Y=MeanA(2:end);
fo = fitoptions('Method','NonlinearLeastSquares',...
'Lower',[500,0,-100],...
'Upper',[9000,30,400],...
'StartPoint',[1528,15, 90]);
ft = fittype('A*exp(-R2*x)+B','options',fo,'coefficients',{'A','R2','B'});
f2= fit(X',Y',ft)
The result I get from the fit is:
f1 =
General model:
f2(x) = A*exp(-R2*x)+B
Coefficients (with 95% confidence bounds):
A = 1597 (1551, 1643)
R2 = 15.64 (15.01, 16.27)
B = 97.78 (89.36, 106.2)
But I don't know how to calculate the standard error of the coefficients (A, B and R2) from the confidence intervals? I know that I can find the standard error of the coefficients using other fit functions such as fitnlm but I need to find the standard error of the coefficients using the "fit" function because I need to apply Lower and Upper bounds.
Thanks very much!
0 Comments
Answers (0)
See Also
Categories
Find more on Linear and Nonlinear Regression 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!