rsquared

versión 1.0.0.0 (2.05 KB) por R P
calculate standard and adjusted R-squared (coefficient of determination)

741 descargas

Actualizada 5 Dec 2016

Ver licencia

rsquared calculates the coefficient of determination (r2) from the original
data (ydata) and fited data (yestimation). It also calculates the adjusted
coefficient (r2adj) considering the number of parameters of the model
(nparam).

Syntax:
r2 = rsquared(ydata,yestimation)
[r2,r2adj]=rsquared(ydata,yestimation,nparam)

Example:
xdata = [1 5 14 23 25 48 49 59 73 77 99 ];
ydata = [-100 70 100 450 550 2200 2300 3400 5300 5906 9600];
plot(xdata,ydata,'ok'), hold on
param_1 = polyfit(xdata,ydata,1);
yestimation_1 = polyval(param_1,xdata);
[r2_1,r2adj_1]=rsquared(ydata,yestimation_1,length(param_1))
plot(xdata,yestimation_1,'-r')
param_2 = polyfit(xdata,ydata,2);
yestimation_2 = polyval(param_2,xdata);
plot(xdata,yestimation_2,'-b')
[r2_2,r2adj_2]=rsquared(ydata,yestimation_2,length(param_2))
legend({'data',['r2=' num2str(r2_1) ', r2adj=' ...
num2str(r2adj_1)],['r2=' num2str(r2_2) ', r2adj=' num2str(r2adj_2)]}, ...
'Location','best')

Equations
SSres=sum( (ydata-yestimation).^2 ); % residual sum of squares
SStot=sum( (ydata-mean(ydata)).^2 ); % total sum of squares
r2=1-SSres/SStot; % standard rsquared
r2adj = 1 - SSres/SStot * (length(ydata)-1)/(length(ydata)-nparam); % adjust for the number of parameters

Check https://en.wikipedia.org/wiki/Coefficient_of_determination

Citar como

R P (2022). rsquared (https://www.mathworks.com/matlabcentral/fileexchange/60577-rsquared), MATLAB Central File Exchange. Recuperado .

Compatibilidad con la versión de MATLAB
Se creó con R2016a
Compatible con cualquier versión
Compatibilidad con las plataformas
Windows macOS Linux
Etiquetas Añadir etiquetas

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!