# nlinfit fitting error approaching infinity

2 views (last 30 days)
Answered: Deepak Meena on 28 Feb 2021
I am trying to run the code below:
However I am getting the error:
Error using nlinfit>checkFunVals (line 649)
The function you provided as the MODELFUN input has returned Inf or NaN values.
Error in nlinfit>LMfit (line 620)
if funValCheck && ~isfinite(sse), checkFunVals(r); end
Error in nlinfit (line 284)
[beta,J,~,cause,fullr] = LMfit(X,yw, modelw,beta,options,verbose,maxiter);
How can I solve this issue
A(:,:,1) = [1.17 1.2 1.22 1.23 1.25 1.27
1.18 1.21 1.22 1.23 1.25 1.25
1.19 1.21 1.22 1.23 1.24 1.24
1.2 1.22 1.22 1.22 1.23 1.22
1.2 1.21 1.22 1.22 1.22 1.22
1.2 1.21 1.21 1.21 1.21 1.2
1.2 1.2 1.2 1.2 1.2 1.19];
A(:,:,2)=[ 1.22 1.26 1.31 1.33 1.37 1.39
1.26 1.29 1.32 1.33 1.36 1.37
1.27 1.3 1.32 1.33 1.36 1.36
1.28 1.31 1.33 1.34 1.35 1.34
1.29 1.31 1.32 1.33 1.34 1.33
1.29 1.3 1.31 1.32 1.32 1.31
1.29 1.3 1.31 1.31 1.31 1.29];
A(:,:,3) = [1.29 1.34 1.37 1.41 1.45 1.47
1.31 1.36 1.4 1.42 1.44 1.45
1.33 1.38 1.41 1.41 1.43 1.43
1.35 1.39 1.41 1.41 1.42 1.41
1.35 1.39 1.41 1.41 1.41 1.4
1.36 1.38 1.4 1.39 1.4 1.38
1.36 1.38 1.39 1.39 1.39 1.37];
K = 60:10:120;
C = [0 0.1 0.2 0.3 0.6 0.9];
X = [7 10 13.5].^2/100^2;
[c, k, x] = meshgrid(C, K, X);
% % % MAPPING: x = ckx(:,1), y = ckx(:,2), z = ckx(:,3), a = b(1), b= B(2), c = b(3), d = b(4)
ckx = [c(:) k(:) x(:)];
Eq = @(b,ckx) ((b(8).*ckx(:,2).*ckx(:,1)) + 1).*(b(1).*ckx(:,2) + b(2).*(ckx(:,1)).^b(3)).* ckx(:,3).^(b(4).*ckx(:,2)+b(5).*(ckx(:,1)+1).^b(6)) + b(7);
B = nlinfit(ckx, A(:), Eq, [0.01, 1, 1, 0.01, 0.1, 0.1, 0.1, 0.00001]);
plot(A(:)); hold on; plot(A(:)-R)
xticks([4:7:126])
xticklabels([0 0.1 0.2 0.3 0.6 0.9 0 0.1 0.2 0.3 0.6 0.9 0 0.1 0.2 0.3 0.6 0.9])

Deepak Meena on 28 Feb 2021
From my understanding , your initial values of the parameters seems to be problem since you are using exponential terms
Using this answer, I changed the initial values to
[0.01, 1, 0.09, 0.0018, 0.1, 0.19, 0.1, 0.00001]
It doesn't give the Nan or Inf error but this might not be the best starting values you neeeded for your dataset.