How can calculate R-Square using lsqcurvefit

Dear Sir/Madam,

I would appreciate how can calculate R-Square using lsqcurvefit or how can calculate this parameter between the real profile and adjusted profiles.

Thanks

Regards

Answer 1

[Star Strider]

You have to calculate it yourself, however the calculation (for vector independent and vector dependent variables) is straightforward:

SStot = sum((y-mean(y)).^2);                            % Total Sum-Of-Squares
SSres = sum((y(:)-yfit(:)).^2);                         % Residual Sum-Of-Squares
Rsq = 1-SSres/SStot;                                    % R^2

Here, ‘y’ is the dependent variable, and ‘yfit’ is the fitted value of the dependent variable from the regression.

[Antonio Gamez]

Also, you can directly write:

R_sq= 1-var(y-yfit)/var(y);

which is equivalent.

[Ha Yin LEE]

Plain and straightforward. Thanks a lot

[Star Strider]

@Ha Yin LEE — My pleasure!