In the Adjusted R-squared you showed hoe to remove the unimportant features but in that example there are two features but what if there are 40 features it would be very difficult to see the hypothesis and rejecting the features how to deal if there are more features
Thanks & Regards,
You are correct that with 40 features it is very difficult. That is one of the jobs a data scientist should do – take the time to go through different combinations of features.
However, also not that 40 features are too many to be completely different. It is quite possible that you can combine some of them, remove others, because you know right away they are not useful, make dummies from the categorical ones.
Finally, you’d first consult the p-value of the coefficient and only then look at the change the Adjusted R-squred.