Fitting the data without using curve fitting tool

Hi. I have a matrix and I have been using the following syntax to fit some data from the matrix z.

[~,~,output] = fit(t,z(:,32),'poly3');

p= output.residuals

The output residuals get stored as output.residuals in the workspace and have to be extracted.

I wish to do the fitting for all the columns in the matrix and use the residuals for all columns in a loop. I wish to put the fitting syntax directly in the loop and not extract output residuals since the loop doesn't run.

[Matt J]

sagi moshe lifshitz's comment moved here:

I have a follow-up question:

I also used the fit() function (on 2017a):

[ fitobject, gof, output ] = fit( [x, y], z, 'poly11' );

plot( fitobject, [x, y], z );

Now in 2020b it isn't working,

The defult function is fit() from machine.learning.toolbox (https://www.mathworks.com/help/stats/lime.fit.html),

and not the regular fit() from curve.fitting.toolbox (https://www.mathworks.com/help/curvefit/fit.html).

The only way to use it is to buy the toolbox? :(

Answer 1

[KSSV]

YOu can use polyfit.

p = polyfit(t,z(:,32),3) ; 
zi = polyval(p,t) ;