Weighted precision & recall - With class weights vs oversampling

Asked Jan 06 '23 at 03:53

Active Jan 06 '23 at 04:00

Viewed 21 times

If weighted precision is calculated as

weighted_precision = [(pc1*c1)+(pc2*c2)]/(c1+c2)
where,
pc1 = precision of class 1
c1 = no.of samples in class 1
pc2 - precision of class 2
c2 = no.of samples in class 2

How does the above equation change while dealing with class imbalance via

Class Weights

Oversampling

edited Jan 06 '23 at 04:00

asked Jan 06 '23 at 03:53

Kaushik J

I am dealing with class imbalance problem, will the weighted precision be different if I use class_weights vs oversampling? – Kaushik J Jan 06 '23 at 03:57
Have you read this thread Are unbalanced datasets problematic, and (how) does oversampling (purport to) help?? – dipetkov Jan 06 '23 at 08:24

Weighted precision & recall - With class weights vs oversampling

0 Answers0