This question is closely similar, however, the answer only speaks about Logistic Regression being one example.
I am interested in knowing if there are more algorithms that are not affected (at least lesser affected than other algorithms) by datasets that have class imbalance.
Asked
Active
Viewed 64 times
1
-
1Welcome to Cross Validated! What issue(s) do you see class imbalance causing? – Dave Apr 26 '22 at 19:54
-
Thanks, Dave. The most common ones would be poor performance and bias. – Rajdeep Biswas Apr 26 '22 at 20:01
-
3Bias (what precisely do you mean by "poor performance"? An accuracy of 99.8% does not sound "poor", until one thinks about the class imbalance more deeply) arises not because of unsuitable algorithms, but because of evaluation metrics that reward it. Are unbalanced datasets problematic, and (how) does oversampling (purport to) help? – Stephan Kolassa Apr 26 '22 at 20:13