Polyfit Ignoring NaN Values

Hello!

I am trying to iterate over rows of a matrix (Z) and fit a curve to each row and then subtract the fitted curve from the original data, however all of my rows contain NaN values. I would like to somehow fit the curve ignoring the NaN values, to subtract this curve from my original matrix row of values. The result would be a matrix with the same dimensions as my original, with the curve subtracted from real values and NaN values in their original place holders. I have thought about indexing NaN values using isnan() and then applying the polyfit to values when isnan() == 0 but everything I've tried so far hasn't quite worked out.

This is what it looks like so far:

for j = 1:m;
    nx = size(Z,2);
    x = 1:nx;
    p = polyfit(x,Z(j,:),2);
    y = polyval(p,x)
    Z(j,:) = Z(j,:) - y;
end

Thanks for your help!

Answer 1

[Image Analyst]

Can you attach Z in a .mat file so we can try things? Like

for j = 1:m;
    thisRow = Z(j, :);
    nx = size(Z,2);
    x = 1:nx;
    nanIndexes = isnan(thisRow);
    thisRow(nanIndexes) = [];
    x2 = x(~nanIndexes)
    p = polyfit(x2,thisRow,2);
    y = polyval(p,x)
    Z(j,:) = Z(j,:) - y;
end

but that's just off the top of my head. would be easier if we had your Z.

[Maddie Harasyn]

This works perfectly! I tried to attach my matrix but it was deemed too large, as it is a DEM ASCII file of a large drone survey I ran. I think I was having trouble, as I am used to Python and was trying to complete this with embedded for and if loops.

Thanks so much!