Why weighted F1-measure, precision and recall are always very similar to accuracy in my problem?

Question

I'm adopting accuracy and macro and weighted averages of precision, recall and f-measure for evaluating my model in a multiclass problem with an imbalanced dataset.

However, I noticed that weighted f1, precision, and recall are always very similar in this context. Is it common? Is it a good practice only use accuracy and macro averages, without weighted averages? Are there scenarios where is justified using weighted averages of precision, recall and f-measure?

If you post very closely related questions, please link them together, which makes life much easier on later people who are interested in the topic. Your first question was Is accuracy enough for evaluating models in balanced multiclass problems?, your second question was Are precision, recall and f-measure more informative than accuracy in multiclass classification with balanced data? — Stephan Kolassa, Mar 28 '23 at 13:51
Also, please explain how the weighting in your F1 score works, and how precisely you define precision and recall in the multiclass context. Thank you! — Stephan Kolassa, Mar 28 '23 at 13:52
I'm using these definitions of accuracy, precision and recall and macro and weighted averages. https://towardsdatascience.com/micro-macro-weighted-averages-of-f1-score-clearly-explained-b603420b292f — Zaratruta, Mar 28 '23 at 13:59
They are very similar in the sense that sometimes the numbers are exactly the same, and sometimes thy differ only in the second decimal digit. — Zaratruta, Mar 28 '23 at 14:00
I posted as a separate question because for me it is a different question. — Zaratruta, Mar 28 '23 at 14:00
But the question remains...Why accuracy is so similar to weighted f1-measure, precision and recall in this problem? Is it common? — Zaratruta, Mar 28 '23 at 18:50

Why weighted F1-measure, precision and recall are always very similar to accuracy in my problem?

0 Answers0