Error: Error using lsqcurvefit Function value and YDATA sizes are not equal.

Hi all!

I have written the following script (just show the important part):

grad=xydata(:,1);
int=xydata(:,2);

b=(nuclgamma*2*p30)^2

exponent_cnst= (nuclgamma*2*p30)^2 * (d20 -p30/3);

Delta_prime=(d20 -p30/3)

x = linspace(min(grad),max(grad));

options = optimset('MaxFunEvals', 300,'MaxIter', 300,'TolFun',1e-12,'Display','off');

D1_prime= @(x,p) (0.5*(p(1)+p(2)+1./(b*x.^2)*p(3))+sqrt(sqrt(p(1)+p(2)+1./(b*x.^2*p(3)))+4./(b^2*x.^4*p(4))))

D2_prime= @(x,p) (0.5*(p(1)+p(2)+1./(b*x.^2)*p(3))-sqrt(sqrt(p(1)+p(2)+1./(b*x.^2*p(3)))+4./(b^2*x.^4*p(4))))

p2_prime= @(x,p) ((p(5)*p(1)+p(6)*p(2)-D1_prime(x,p))/(D2_prime(x,p)-D1_prime(x,p)))

p1_prime= @(x,p) (1-((p(5)*p(1)+p(6)*p(2)-D1_prime(x,p))/(D2_prime(x,p)-D1_prime(x,p))))

model= @(x,p) p1_prime(x,p)*exp(-exponent_cnst.*D1_prime(x,p).*Delta_prime)+p2_prime(x,p)*exp(-exponent_cnst.*D2_prime(x,p).*Delta_prime)

%              p(1)   p(2)    p(3)        p(4)       p(5)    p(6)            
%              D1      D2  1/tau1+1/tau2  tau1*tau2   P1      P2   
%               

%               p(1)   p(2)   p(3)  p(4)     p(5)  p(6)  
startingVals = [1e-9   1e-8   1e5   1e-3     0.5   0.5   ]; %guess values
lb =           [1e-13  1e-13  1      1e-12     0    0   ]; %lower bound
ub =           [1e-4   1e-4   1e10   1e-6     1     1  ]; %upper bound

[p] = lsqcurvefit(model,startingVals,grad,int,lb,ub,options);

I've checked the script again and again but don't find my mistake :-( I'm very new with MATLAB so please forgive if my question is kind of trivial.

Many thx in advance!

Answer 1

[Alan Weiss]

MATLAB think that your objective function is not giving you a return vector of the correct size. Take your x data, grad, take a vector x of potential parameter values (might as well take lb or ub), and evaluate

 model(x,grad)

See if you get a result the same size as your y data, int.

If not, then fix things until you do.

If so, then please post the entire error message, copy and paste, do not retype.

Alan Weiss

MATLAB mathematical toolbox documentation

Answer 2

[Torsten]

You switched the order of the arguments in "model": use model(p,x) instead of model(x,p).

Furthermore, the line

x = linspace(min(grad),max(grad));

can be deleted.

Best wishes

Torsten.