Surface fit doesnt cover all datapoints

I am triying to plot a surface response with some input points from a parametric study with the following code:

[xData,yData,zData] = prepareSurfaceData(x,y,z);
ft = 'cubicinterp';
[fitresult,gof] = fit([xData,yData], zData, ft, 'Normalize', 'on');
[X,Y] = meshgrid(x(1):0.1:xData(end),y(1):0.1:y(end));
Z = fitresult(X,Y);
surf(X,Y,Z,'EdgeColor','texturemap');

The problem is that the surface fitted doesn't cover all original points x,y,z (red balls in figures), it stops in a given x () value. I am fitting ten different surfaces with the same x and y data points but different z and all have the same problem.

[dpb]

Attach at least one of the datasets -- nobody can reproduce anything easily.

It would appear that the problem is probably that your [X,Y] vectors don't span the space; you use limits from the first and last elements in an input array, but those apparently aren't necessarilyh the minimum nor maximum points in the datasets.

[Jesús Gutiérrez]

Thanks for the response! Just attached two datasets I'm using for the surface. (I know there are very few datapoints for a proper surface fit but its all I have available).

Answer 1

[John D'Errico]

load dataset1
x
x =
    20    30    40    50    20    30    50    40    50    60    40    50    60    20    30    50    60    70    30    40    50

Why am I surprised that x(1) and x(end) do not contain all the data?

x(end)
ans =
    50
max(x)
ans =
    70

How did you display the data?

[X,Y] = meshgrid(x(1):0.1:xData(end),y(1):0.1:y(end));

Of course, this line of code does not even use x(end), but xData(end).

[xData,yData,zData] = prepareSurfaceData(x,y,diff_fpant_1_max);
>> xData
xData =
    20
    30
    40
    50
    20
    30
    50
    40
    50
    60
    40
    50
    60
    20
    30
    50
    60
    70
    30
    40
    50

However, does prepareSurfaceData actually perform a sort of your data? It might be a good idea to look at that.

I would suggest you are jumping to conclusions without actually looking at what some of that code does. That would alsys seem a dangerous thing to do.

[Jesús Gutiérrez]

Thanks for your help, I wasn't sure where to start finding the problem. I got this code from the internet but I didn't understand it well... I just introduced min(x), max(x), min(y), max (y) as arguments instead of (1) and (end) and now i get a surface that at least covers all input points.