MATLAB: Is it possible to use a for loop to change which linear regression to use for optimization purposes

fitlinear regressionoptimizationrsquare

I'm trying to optimize my code where I want to discover which linear regression model is optimal to use, looking at R^2. So far this is what I've come up with (where M1 and M2 are parts of two tables):
[linear1, gof]=fit(M1,M2,'poly1');
linear1_R2=gof.rsquare
% 

[linear2, gof]=fit(M1,M2,'poly2');
linear2_R2=gof.rsquare
% 
[linear3, gof]=fit(M1,M2,'poly3');
linear3_R2=gof.rsquare
...
I want to do this until poly9 and it feels like it can be optimized. I was thinking about doing a for loop where the program (1) will run all the fit functions and (2) display which one gives the highest R^2 and what that value is, but I'm not sure how to proceed.
Hope someone can help, thanks in advance!

Best Answer

Hi Anna,
The below code snippet can accomplish the task. It fits all the models specified it 'fit_func' array and records the R^2 value in a vector. Finally, the model with the highest R^2 value can be easily extracted.
fit_func={'poly1','poly2','poly3','poly4','poly5','poly6','poly7','poly8','poly9'};
rsquare_values=zeros(length(fit_func),1);
for i=1:length(fit_func)
    [linear1, gof]=fit(M1,M2,fit_func{i});
    rsquare_values(i)=gof.rsquare;
end
[best_rsquare,index] = max(rsquare_values);
best_model = fit_func(index);
best_rsquare - highest R^2 value corresponding to the best model.
best_model - contains the model with the highest R^2 value.
Related Question