0
$\begingroup$

The difference between the two models is the label, which is slightly differently defined for the 2nd model. So for model 2 there are less data that fall in class 1.0 (87161 versus 155915).

Which of the two models would you prefer? how should I evaluate a higher f1 score for the separate classes?

Model 1

              precision    recall  f1-score   support

        0.0       0.69      0.90      0.78    299032

        1.0       0.56      0.23      0.32    155915

avg / total       0.65      0.67      0.63    454947

Model 2

              precision    recall  f1-score   support

        0.0       0.87      0.93      0.90    367786

        1.0       0.58      0.40      0.47     87161

avg / total       0.81      0.83      0.82    454947
$\endgroup$
-1
$\begingroup$

It seems that by moving from 1st model to the second, every individual metric is enhanced, so I would say the 2nd model is definitely better.

| cite | improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.