how to obtain coefficients and other data from a curve fit object, of the interpolant type, using linear method?

From a set of data, I could easily obtain a curve fit using curveFitter. The fit is of the interpolant type, using linear method. From what I understand, the underlying method is linear triangular interpolation. But then, I'm unable to get the internal data of the model. I mean, there should be triangle boundaries, coefficients and other constants describing the triangular pieces of surface that approximate my data set, but I was unable to read them from the obtained fit.

So, for this particular fit (linear interpolant), how to get the inside data?

[Steven Lord]

What are you hoping to do with the internal implementation details of the fit? Are you hoping to use those to evaluate the fit? If so see the functions and topics on the Fit Postprocessing documentation page.

[Fabrice Belveze]

Hello, I did this fit in the perspective of a harware implementation, therefore I have to precisely know how it works.

I did use the functions you sent me the link to, unfortunately, they return information that I can't use. For instance, "coeffvalues" returns:

ans =

TriangleSurfaceInterpolant with properties:

Type: 'Linear'

which does not give any numerical value, nor any insight on how the fitted object works.

More precisely, I'm not interested in the method which obtains the fit. I'm interested in how the fitted object performs surface approximation.

[dpb]

If use the fitted object as it is documented to be used, it most certainly does return values--

load census               % get some data to use
format bank               % make it print in %f form  for legibility
[cdate(1:5) pop(1:5)]     % look at first few elements

ans = 5x2

1.0e+00 * 1790.00 3.90 1800.00 5.30 1810.00 7.20 1820.00 9.60 1830.00 12.90

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

f=fittype('poly2');             % define a quadratic model
c=fit(cdate,pop,f);             % perform the fit
format shorte, format compact   % coefficients vary a lot in magnitude so we can see them all
coeffs=coeffvalues(c)           % retrieve the coefficients

coeffs = 1x3

1.0e+00 * 6.5411e-03 -2.3510e+01 2.1130e+04

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

phat=c(cdate(1:5))                   % evaluate the fit over same first few points

phat = 5x1

5.5830e+00 5.3121e+00 6.3495e+00 8.6951e+00 1.2349e+01

<mw-icon class=""></mw-icon>

<mw-icon class=""></mw-icon>

plot(cdate,pop)                      % all the data
hold on
plot(cdate(1:5),phat,'r-')           % add the fitted

figure
plot(c)                 % you get the whole thing automagically if just use the object

In short, read the full documentation and look at the example usage before claiming something doesn't work.

However, for use on an embedded system or the like, you'll undoubtedly be far better served to not use such high level MATLAB features; particularly if all you're after is linear interpolation. You can code that in very few lines much more efficiently. In addition, I suspect a fair amount of this may not be available for the environment you're looking at.