How to Fit Tabular Data

I have a table with two variables, x and y. For each x,y pair there is a z.

I'm trying to find a two variable function that can represent this table. I have been trying to use fit() but it says that the format of my vectors is off. So I did the example and looked at the three vectors loaded in.

load franke;

surffit = fit([x,y],z,'poly23','normalize','on')

I find that the vectors , x, y, and z are all the same dimension. However, in my mind z = f(x,y) should be a matrix of values with dimensions size(x) by size(y). What is this fit function actually fitting then?

[Christopher Dottellis]

To clarify, this data is sparse. Might it just be easier to create the function by hand?

Answer 1

[Nitin Kapgate]

'z' vector should be of the same size as 'x' and 'y' vectors. The fit function tries to approximately fit a curve or surface to the data using different fitting techniques (fit type) etc.

You can refer to this example that demonstrates how to fit a surface using variables in a MATLAB Table.