Can I use lsqcurvefit for a function of multiple variables?

Here is the code I am trying to run:

coeff0 = [1 0.5 2];
svData(:,1) = varStruct(indVar).sv1vals'
svData(:,2) = varStruct(indVar).sv2vals'
zData = varStruct(indVar).linearB'

b = lsqcurvefit(klaCorr,coeff0,svData,zData);

The function klaCorr, which I have in a separate function m-file:

function klaX = klaCorr(coeff,sv)
    klaX = coeff(1) + (sv(1)^coeff(2))*(sv(2)^coeff(3));
end

And the output:

svData =

  248.8431    1.7500
  400.4011    1.7500
  400.4011    2.0000
  103.7757    1.7500
  103.7757    1.7500
  103.7757    1.7500


zData =

   1.0e-04 *

    0.0637
    0.1115
    0.1272
    0.0468
    0.0504
    0.0321

Not enough input arguments.

Error in klaCorr (line 2)
    klaX = coeff(1) + (sv(1)^coeff(2))*(sv(2)^coeff(3));

Error in klascripts (line 609)
b = lsqcurvefit(klaCorr,coeff0,svData,zData);

Does anyone know why I am getting this error message?

Also, if there is a better way to do this with a function other than lsqcurvefit, I am open to suggestions!

Feel free to ask for more details :)

Thanks

Answer 1

[Star Strider]

Your objective function ‘klaCorr’ needs to address the appropriate columns of ‘sv’:

klaCorr = @(coeff,sv) coeff(1) + (sv(:,1).^coeff(2)).*(sv(:,2).^coeff(3));

I used an anonymous function for it here.

The entire code:

svData = [248.8431    1.7500
          400.4011    1.7500
          400.4011    2.0000
          103.7757    1.7500
          103.7757    1.7500
          103.7757    1.7500];

zData = 1.0e-04 * [0.0637
                   0.1115
                   0.1272
                   0.0468
                   0.0504
                   0.0321];

coeff0 = [1 0.5 2];
klaCorr = @(coeff,sv) coeff(1) + (sv(:,1).^coeff(2)).*(sv(:,2).^coeff(3));
b = lsqcurvefit(klaCorr,coeff0,svData,zData)

produces:

b =
         -1.00002742240113      4.42835377098216e-06      1.96276255078078e-05

[Rebecca Wenokor]

Thank you! It works now

[Star Strider]

My pleasure!

If my Answer helped you solve your problem, please Accept it!