I analyzed a dataset using those 3 different algorithms. As I can see, Random forest performs better in most cases. My dataset is composed of 4000 instances of two classes (class A 2000 elements, class B 2000 elements). I use 207 metrics to classify the instances, but I also use the first 20 or 10 best metrics for InformationGain. My question is: why sometimes an algorithm performs better than another one (in this case I'm only comparing this 3). I read about them but I would like to have a complete scenario of why in some case RF is better than Bayes net and why sometimes is the opposite. And why SMO is always worst than the other two, in my experiences. Thank you so much!
1 Answer
In machine learning, there's often no clear "why" or "reason" we can point to. These are different methods, which implicitly embed different biases or different priors or different assumptions, and thus will work better in different situations.
-
$\begingroup$ Ok, but do you know possible reasons for the better performance of an algorithm instead of another one? $\endgroup$ Nov 22, 2020 at 13:40