error using curve fitting tools--Attempted to access x(8); index out of bounds because numel(x)=7

Hi I am a newbie in matlab, I want to use curve fitting tool to fit a custom equation, which is a continuous two-peak curves, the model includes 8 parameters.

function y = minimize(x,a,b,c,d,e,f,g,h)
TC = xlsread('M:\G1-L1-D1-1.xls',1,'I3:I258');
R2 = 1000000./ TC;
nrows = size(R2,1);
for i = 1:nrows
    if (x(i)<a)
        y(i)=0;
    else if (a<x(i)<b)
          y(i)= h.*d.*(sin(pi.*(x(i)-a)./(b...
    -a))).*exp(-(c.*(x(i)-a)))./((b-a).*pi.*(1+exp(-(c.*(b-a...
    )))./((c.*(b-a))^2)+pi.*pi));
     else if (x(i)>b)  
         y(i)=0;
     else if (b<x(i)<e)
                   y(i)=0;
     else if (e<x(i)<f)
                        y(i)= h.*(1-d).*(sin(pi.*(x(i)-e)./(f-e))).*exp(-(g.*(x(i)-e)))./((f-e).*pi.*(1+exp...
    (-(g.*(f-e))))./((g.*(f-e))^2+pi.*pi))
     else if (x(i)>f)
                       y(i)=0;

           end
           end
           end
           end
          end
      end
      i=i+1;
  end

And then I call cftool, but got an error like this, I don't understand why and how can I solve this, is it something to do with start points? Thanks! Expression minimize(x,a,b,c,d,e,f,g,h) is not a valid MATLAB expression, has non-scalar coefficients, or cannot be evaluated: Error in fittype expression ==> minimize(x,a,b,c,d,e,f,g,h) ??? Attempted to access x(8); index out of bounds because numel(x)=7.

close all
clear all
TC = xlsread('M:\G1-L1-D1-1.xls',1,'I3:I258');
x = 1000000./ TC;
x = x(98:198,1);
QQ = 'M:\ILT-D1';
aa = dir(QQ);
aa = xlsread('M:\G1-L1-D1-1.xls',1,'J3:J258');
y= aa.* TC;
y = y(98:198,1);
cftool

regards Han

Answer 1

[Walter Roberson]

Your code

 if a<x(i)<b

means

if (a<x(i)) < b

The (a<x(i)) part evaluates to either 0 (false) or 1 (true), so you are then comparing 0 or 1 to b. It is highly unlikely that you want to do that.

Your code reads I3:I258 of the xls file. That is 256 entries. That number, 256, gets used as nrows, which is the maximum index in "for i". You index x(i) . Therefore your input, x, must have 256 entries in the vector, but instead it has 7. I am not certain at the moment why it would have 7 entries considering that you use x = x(98:198,1) which should have 11 entries. Either way, 7 entries or 11 entries is certainly not the 256 that you are looping over.

Notice also that if

if (x(i)>b)

is true whenever b <= x(i) . And then in your next test you check

if (b<x(i)<e)

but if b<x(i) then the previous test would have been true.

I recommend also that you program using "elseif" rather than "else if". And remember that any test that has succeeded earlier does not need to be repeated, so if you tested x <= a then the next test does not have to check a < x because you could not have reached the second "if" in the situation where x <= a

[Han Zhu]

Thank you for your answer. The function that I want to fit to depends on the range of x. I correct it as this:

function y = minimize(x,a,b,c,d,e,f,g,h)
TC = xlsread('M:\G1-L1-D1-1.xls',1,'I3:I258');
R2 = 1000000./ TC;
nrows = size(R2(98:198,1),1);
for i = 1:nrows

      if (x(i)>a && x(i)<b)
            y(i)= h.*d.*(sin(pi.*(x(i)-a)./(b...
      -a))).*exp(-(c.*(x(i)-a)))./((b-a).*pi.*(1+exp(-(c.*(b-a...
      ))))./((c.*(b-a))^2+pi.*pi));
      elseif (x(i)>e &&x(i)<f)
            y(i)= h.*(1-d).*(sin(pi.*(x(i)-e)./(f-e))).*exp(-(g.*(x(i)-e)))./((f-e).*pi.*(1+exp...
      (-(g.*(f-e))))./((g.*(f-e))^2+pi.*pi));
       else
                         y(i)=0;

      end
  end
  end

I don't really understand your comment " I am not certain at the moment why it would have 7 entries considering that you use x = x(98:198,1) which should have 11 entries." why is it 11 entries? isn't it 101?

I tried to run it again, but the same error message still appears.. Also I tried the fittype function instead of cftool:

ft = fittype('minimize(x,a,b,c,d,e,f,g,h)');
f = fit( x, y, ft,'StartPoint', [3.410941423,5.321103981,0.609720564,0.015015503,10.41287485,43.70890568,0.078584389,1.04E+09]);

which gives error as Error using fittype>iTestCustomModelEvaluation (line 726) Expression minimize(x,a,b,c,d,e,f,g,h) is not a valid MATLAB expression, has non-scalar coefficients, or cannot be evaluated: Error in fittype expression ==> minimize(x,a,b,c,d,e,f,g,h) ??? Undefined function 'minimize' for input arguments of type 'double'.

Error in fittype>iCreateFittype (line 367) iTestCustomModelEvaluation( obj );

Error in fittype (line 324) obj = iCreateFittype( obj, varargin{:} );

Caused by: Error using fittype/evaluate (line 102) Error in fittype expression ==> minimize(x,a,b,c,d,e,f,g,h) ??? Undefined function 'minimize' for input arguments of type 'double'.

[Walter Roberson]

101,yes,I misread.

In order to use fittype with a string instead of an anonymous function, the function named in the string must have its own .m file and cannot be a function defined in some other .m file.