Inveresely evaluate a sFit Object

I have the following for-loop

for i=0:0.001:4000;
   if feval(sFit,2,i) <= 0.9
         currentY = i;
         break;
    end
end
returnValue =  feval(sFit,2,currentY+20);

I'm evaluating an sFit-Object inversely for a z = 0.9 and x = 2 to get the y-value. (The sFit is monotonically decreasing with x and y). After the Loop, I evaluate the sFit-object "conventially" with X and Y to get Z as a result.

Unfortunately this takes forever to compute, as the feval-function is called many times.

Is there any way to inversely evaluate a sfit-object quicker? I want to get the y-value for given z and x-values immediately, without any loop.

[José-Luis]

What is it you are trying to achieve?

[Andreas Harter]

The z-values range from 0 to 1 and describe a degeneration process. Depending on the input values of x and y the degeneration is stronger or weaker. The sFit-Object is a surface-Fit of Lab-Data that I use as a look-up-table for my real-time-inputs of x and y.

Answer 1

[Steven Lord]

Use fzero. In this example, when fzero finds a zero of myfun that value of y will satisfy sf(0.5, y) - 1 = 0 or sf(0.5, y) = 1.

% Sample data
[x, y, z] = peaks;

% Perturb the Z data slightly
z = z + rand(size(z));

% Create a surface fit object
sf = fit([x(:) y(:)], z(:), 'poly22');

% Create the function on which fzero will operate
myfun = @(y) sf(0.5, y)-1;

% Call fzero with an initial guess of y = 2.
desiredY = fzero(myfun, 2);

% Check by evaluating the fit
shouldBeEqualTo1 = sf(0.5, desiredY)

[Andreas Harter]

Thank you, that did the trick!