Adding Function Handles with "different" parameters and different factors?!

Hi there,

I have smth. like this:

%some data i want to fit
xfit = somedata;
yfit = somedata;
% init handles?
summed_EOF1 = @(x) 0;
EOF1 = @(x) 0;
helperfun = @(x) 0;
for i = 1 : length(xfit)
    %before it will be decided how many terms are used, e.g. 3
    for j=length(number_of_terms):-1:1
        if j == 1
            helperfun = @(x) (helperfun(x) + abs(x(1)*xfit(i) + x(2)*yfit(i) + x(3))^x(4));
        else
            helperfun = @(x) (helperfun(x) + abs(x( j * 4 + 1)*xfit(i) + x(j * 4 + 2)*yfit(i) + x(j * 4 + 3))^x(j*4  + 4));
        end
    end
% error of fit function
EOF1 = @(x) (1 - (helperfun(x))^(1/5) )^2;
summed_EOF1 = @(x)( summed_EOF1(x) + EOF1(x));
end

in order to get a cost function, but I am not sure if this code really sums all terms adequately. When I minimize this cost function only the first 4 parameters are fitted and the rest remains at the initial guess.

Does someone know how to properly sum function handles?

Greets,

Pay

[Walter Roberson]

if you have the Symbolic toolbox constuct symbolic formulae and matlabFunction when done to create the function handle .

[Payjay]

Thanks for you're answer! I also considered this, but I did not find a way to "dynamically" generate symbolic variabels, like x(i), do you know a way to do this with symbolic variables? Then I would absolutely do this! :-)

[Walter Roberson]

You do not need to. You know the number of variables ahead of time.

x = sym('x', 1, N)

Then index x as needed .

Be sure to matlabFunction() 'vars', {x}

so that it bundles all the entries into a vector when generating code.

[Payjay]

Many thanks! :)