Doubt in DT quiz

Consider an alternative way of learning a Random Forest where instead of randomly sampling the attributes at each node, we sample a subset of attributes for each tree and build the tree on these features. Would you prefer this method over the original or not, and why?

Yes, because it reduces the correlation between the resultant trees

Yes, because it reduces the time taken to build the trees due to the decrease in the attributes considered

No, because many of the trees will be bad classifiers due to the absence of critical features considered in the construction of some of the trees

Hey @debjanihome, answer is no because many of the trees will be bad classiffiers due to the absence of critical features considered in the construction of some of the trees, reason being that, not all features are important, only few features are important, and while building decission trees these features must be reused again and again to build good decission trees. If we sample them in a fashion which is not random than we will end up with decission trees, which are bad classifiers and hence may ruin the accuracy.

Hope this cleared your doubt.
Plz mark the doubt as resolved in my doubts section . :blush:

1 Like