Error using lsqcurvefit (line 248) Function value and YDATA sizes are not equal.

Hello everybody. I'm trying to find the parameters to best fit a model function having experimental points as xdata and ydata. The programm and functions are:

A=load('data.txt');
xdata=A(:,1);
ydata=A(:,2);
x0=[1 34 10000 0.7]; % initial guess on parameter x
x=lsqcurvefit(@F,x0,xdata,ydata);

Function:

function F = F(x,xdata)
Na=6.0221413e+23;
phi0=0.5;
Rh0=min(xdata);
nius=1.81e-5;
F=(x(1)*x(2)*phi0^2*(Rh0/xdata).^6)/...
    (nius*x(3)/(Na*4*pi*Rh0^3/3)*(1/2*(Rh0/xdata).^3-(Rh0/xdata))-...
    phi0*(Rh0/xdata).^3-log(1-(phi0*(Rh0/xdata).^3))+...
    phi0^2*(x(1)-1/2)*(Rh0/xdata).^6-x(4)*phi0^3*(Rh0/xdata).^9);

I get this two error messages:

Error using lsqcurvefit (line 248) Function value and YDATA sizes are not equal.

Error in datafit (line 12) x=lsqcurvefit(@F,x0,xdata,ydata);

What am I doing wrong? I'm sorry if this is a stupid question, but I'm quite new to Matlab. Thanks.

[Anthony]

Hum... Did you check:

size(ydata)
size(F(x0,xdata))

They should be equal, if not, there is maybe a mistake in your definition of F.

Cheers,

[sumalatha vasam]

size is same for

size(ydata)

size(F(x0,xdata))

Answer 1

[Ben C]

You could have inputs of function reversed. Function inputs should be (x0, x)

Answer 2

[sumalatha vasam]

Facing same problem

Error using lsqcurvefit (line 262)

Function value and YDATA sizes are not equal.

Error in test52 (line 12)

MI=lsqcurvefit(@impedance_test52,[100,4,10e7,2,1,1,10e-6],field,Z,lb,ub);

can u pls help me regarding this issue ....

function modZ=impedance_test52(var,field)

M=var(1);Ha=var(2);sig2=var(3);teta=var(4);si=var(5);alpha=var(6);D=var(7);

%Define constants

g=((2.2).*(10e5));
w=((90).*(10e3));
M0=((4).*(pi).*(10e-7));

l=0.03;

c=(3*(10e8));

% Define function.

wm=g*4*(pi)*M;

w1=g*((Ha*(cosd(teta-si))^2)+(field*(sind(teta))));

w2=g*((Ha*(cosd(2*(teta-si))))+(field*(sind(teta))));

%Relative permeability

Mr=(wm.*(wm+w1-(1i.*(alpha).*w)))/(((wm+w1-(1i.*(alpha).*w)).*(w2-(1i.*(alpha).*w)))-(w^2));

Mr=abs(Mr);

Mef=Mr*M0;

T2=c/(sqrt(2*(pi)*(sig2)*w));

Z=(((1-1i)*l)/((pi)*(sig2)*D*T2))*(((sqrt(Mef+1))*((sind(teta))^2))+(cosd(teta))^2);
modZ=((abs(Z)/(4.4572e-10))-1)*100;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear all;
close all;
load('mat52.txt');
field= mat52(:,1);
Z= mat52(:,2);
 plot(field,Z,'bo');
 
hold on;
lb=[90,2,5.6*(10e7),1,0.1,0.1,0.1*(10e-5)];
ub=[120,15,5.6*(10e8),10,10,5,100*(10e-5)]; 
MI=lsqcurvefit(@impedance_test52,[100,4,10e7,2,1,1,10e-6],field,Z,lb,ub);
fitdata=impedance_test52(MI,field);
plot(field, fitdata,'-b');

xlabel('Field');
ylabel('MI');
legend('mat52','fitting')
title('MI vs Field')
  

xlable and ylable
-15.2	273.20056
-14.4	278.50354
-13.6	283.44331
-12.8	287.36606
-12	291.97893
-11.2	295.96223
-10.4	301.4226
-9.6	305.21823
-8.8	307.87578
-8	310.12168
-7.2	309.65555
-6.4	307.67601
-5.6	309.11677
-4.8	308.11187
-4	302.00375
-3.2	295.5869
-2.4	282.01465
-1.6	266.74738
-0.8	253.90762
0	251.76379
0.8	253.90762
1.6	266.74738
2.4	282.01465
3.2	295.5869
4	302.00375
4.8	308.11187
5.6	309.11677
6.4	307.67601
7.2	309.65555
8	310.12168
8.8	307.87578
9.6	305.21823
10.4	301.4226
11.2	295.96223
12	291.97893
12.8	287.36606
13.6	283.44331
14.4	278.50354
15.2	273.20056
end

[Torsten]

What is

size(Z) 
size(modZ)

?

[sumalatha vasam]

size(Z)

ans =

39 1

>> size(modZ)

Undefined function or variable 'modZ'.

what the problem is?

[Torsten]

function main
  load('mat52.txt');
  field= mat52(:,1);
  Z= mat52(:,2);
  plot(field,Z,'bo');

  hold on;
  lb=[90,2,5.6*(10e7),1,0.1,0.1,0.1*(10e-5)];
  ub=[120,15,5.6*(10e8),10,10,5,100*(10e-5)]; 
  fitdata0 = impedance_test52([100,4,10e7,2,1,1,10e-6],field);
  size(fitdata0)
  size(Z)
  MI=lsqcurvefit(@impedance_test52,[100,4,10e7,2,1,1,10e-6],field,Z,lb,ub);
  fitdata=impedance_test52(MI,field);
  plot(field, fitdata,'-b');

  xlabel('Field');
  ylabel('MI');
  legend('mat52','fitting')
  title('MI vs Field')
end

function modZ=impedance_test52(var,field)
  M=var(1);Ha=var(2);sig2=var(3);teta=var(4);si=var(5);alpha=var(6);D=var(7);

  %Define constants 
  g=((2.2).*(10e5));
  w=((90).*(10e3));
  M0=((4).*(pi).*(10e-7));
  l=0.03;
  c=(3*(10e8));

  % Define function.
  wm=g*4*(pi)*M;
  w1=g*((Ha*(cosd(teta-si))^2)+(field*(sind(teta))));
  w2=g*((Ha*(cosd(2*(teta-si))))+(field*(sind(teta))));

  %Relative permeability
  Mr=(wm.*(wm+w1-(1i.*(alpha).*w)))/(((wm+w1-(1i.*(alpha).*w)).*(w2-(1i.*(alpha).*w)))-(w^2));
  Mr=abs(Mr);
  Mef=Mr*M0;
  T2=c/(sqrt(2*(pi)*(sig2)*w));
  Z=(((1-1i)*l)/((pi)*(sig2)*D*T2))*(((sqrt(Mef+1))*((sind(teta))^2))+(cosd(teta))^2);
  modZ=((abs(Z)/(4.4572e-10))-1)*100;
end

Please save under "main.m", execute and report size(fitdata0) and size(Z). They should be equal.

[sumalatha vasam]

size(fitdata0)

ans =

39 39

>> size(z)

ans =

39 1

so, we have to change Function, Is it right.

[sumalatha vasam]

But how???

[Torsten]

Use

Mr=(wm.*(wm+w1-(1i.*(alpha).*w)))./(((wm+w1-(1i.*(alpha).*w)).*(w2-(1i.*(alpha).*w)))-(w^2));

instead of

Mr=(wm.*(wm+w1-(1i.*(alpha).*w)))/(((wm+w1-(1i.*(alpha).*w)).*(w2-(1i.*(alpha).*w)))-(w^2));

What are the sizes now ?

Answer 3

[sumalatha vasam]

Thank u very much, It's working.

Answer 4

[Joanalyn]

Facing same problem

In three_system (line 53)

Error using lsqcurvefit

Function value and YDATA sizes are not equal.

Error in three_system (line 53)

[theta,Rsdnrm,Rsd,ExFlg,OptmInfo,Lmda,Jmat]=lsqcurvefit(@kinetics,theta0,t,c);

can u pls help me regarding this issue ....

three_system()

Warning: Failure at t=1.452557e+00. Unable to meet integration tolerances without reducing the step size below the smallest value allowed (3.552714e-15) at time t.

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

Error in solution>three_system (line 39)
[theta,Rsdnrm,Rsd,ExFlg,OptmInfo,Lmda,Jmat]=lsqcurvefit(@kinetics,theta0,t,c);

function three_system 
% 2016 12 03
% NOTES: 
% 
%   1. The ‘theta’ (parameter) argument has to be first in your 
%       ‘kinetics’ funciton,
%   2. You need to return ALL the values from ‘DifEq’ since you are fitting
%       all the values

t=[1
2
3
4
5
6
7
8
9
10
11
12];

c=[96   15  4
   98   20  7
   85   18  3
   91   20  4
   75   35  9
   74   37  10
   84   19  18
   73   25  23
   72   26  27
   99   21  24
   102  20  26
   110  18  25]; 
theta0=[1;1;1;1;1;1;1;1;1;1;1];

[theta,Rsdnrm,Rsd,ExFlg,OptmInfo,Lmda,Jmat]=lsqcurvefit(@kinetics,theta0,t,c);

fprintf(1,'\tRate Constants:\n')
for k1 = 1:length(theta)
    fprintf(1, '\t\tTheta(%d) = %8.5f\n', k1, theta(k1))
end

tv = linspace(min(t), max(t));
Cfit = kinetics(theta, tv);

figure(1)
plot(t, c, 'p')
hold on
hlp = plot(tv, Cfit);
hold off
grid
xlabel('Time')
ylabel('Number')
legend(hlp, 'C_1(t)', 'C_2(t)', 'C_3(t)', 'Location','N')

end

function C=kinetics(theta,t)
c0=[100;10;2];
[T,Cv]=ode15s(@DifEq,t,c0);
%
    function dC=DifEq(t,c)
    dcdt=zeros(3,1);
    dcdt(1)= theta(1).*c(1)-theta(4).*c(1)^2-((theta(7).*(c(1)^2).*c(2))\(theta(5)+(c(1)^2)))-((theta(11).*theta(7).*c(1)^2.*c(3))\(theta(5)+c(1)^2));
    dcdt(2)= theta(2).*c(2)-((theta(8).*c(2).*(c(2)+c(3)))\(theta(6)+c(1)))-theta(10).*c(2).*c(3);
    dcdt(3)= theta(10).*c(2).*c(3)+theta(3).*c(3)-((theta(9).*c(3).*(c(2)+c(3)))\(theta(6)+c(1)));
    dC=dcdt;
    end
C=Cv;
end

[Torsten]

Your integration with ode45 is not successful (see above). Thus the returned Cv vector is incomplete (only 1x3 instead of 12x3).

Answer 5

[Tobias Tietz]

Error using lsqcurvefit

Function value and YDATA sizes are not equal.

y = [31.6968154531912	0.890931218661132
    38.4804596257648	0.892775450209936
    46.4695759506599	0.894167809732286
    55.8756531897544	0.895239085956052
    66.9438897413054	0.896075757306964
    79.9650953621297	0.896738163648594
    95.2780467115373	0.897268894599403
    113.280906840510	0.897698637646080
    134.425458339456	0.898049290870298
    159.262240551692	0.898337962105557
    188.417221640351	0.898577105090496
    222.621078145907	0.898776292802748
    262.722843413351	0.898943093784409];
T = [8.82800139811669
    9.40413871093602
    9.96672666696677
    10.5200423966807
    11.0668837672769
    11.6100076414882
    12.1513939504397
    12.6928526284094
    13.2343009424779
    13.7776845090378
    14.3229791441139
    14.8700715607888
    15.4185722748141];
A = [log(y(:,1)),ones(size(T,1),1),1./y(:,2)];
b = T;
c1 = A\b

c1 = 3×1

3.2655 -42.1436 35.3726

c = [1,1,1];
f = @(c, y) c(1).* log(y(:,1)) + c(2) + c(3).*(1./y(:,2));
c2    = lsqcurvefit(f, c, y, T)

Local minimum found.

Optimization completed because the size of the gradient is less than
the value of the optimality tolerance.

c2 = 1×3

3.2655 -42.1436 35.3726

Can someone help me?

[Torsten]

Corrected (see above).

Your problem is linear in the fitting coefficients. It can easily be solved using backslash as added to your code.

A = [log(y(:,1)),ones(size(T,1),1),1./y(:,2)];
b = T;
c1 = A\b