Hi, in the link you provide, an alaytical expression is evaluated on a given grid. Then, using the scatteredInterpolant function, a interpolation is computed. Finally, this interpolation is evaluated on a new grid. I think you can adapt this easily to your problem.

Hi,

I assume you have f( x, y) function to produce the prepared graph. You can use these data to generate interpolate using the function scatteredInterpolant but wirh only 2D data and not 3D data.

Then, you can use extra range for extrapolation.

Thank you for all your answers. In the website I introduced previously I see a V function which was defined in some sample points. My current problem is that I don't have any information about the probable function I may define in sample points.

I would like to know how I can define a function formula like V or f(x,y)?

I would be grateful if you could kindly let me know your viewpoints about that also.

All the best,

Sahar

Hi,

You don't need to have a function because you already have the data and you plot the surface based on your data. In the example you mentioned, it uses the parabolic function to generate the Sample function data.

Right. I finally found the solution considering your guideline. Thank you very much.

Here is the result of my extrapolation. Although it is not as acceptable as I was thought but may be this is the best output I could get.

It does not sound very good extrapolation. This is, I guess, due to your original data is not distributed equally over the peak of surface. If you can manage to get more data around zero point of X axis( negative part), I think you will have better extrapolation.

Yes. Exactly. The problem is that my data points are not available in the negative part of the curve. Since it was a mechanical limitation of my instrument.

However. I am thinking of another solution.

I have data points with lower resolution in the whole part of the curve. and the previous data was with higher resolution in only some parts.

I would like to merge these two surfaces together somehow I will have high resolution information in central part and lower resolution in data in the marginal parts. I would appreciate if you could let me know whether you have any viewpoint about that also or not?

Thank you for your consideration and time.

Here is a picture from my lower resolution data and the interpolated surface on it.

Hi,

I think this is very good idea to use both sets of data with different resolutions. However, I guess the extrapolation will rely mainly on marginal part.