Firstly, is there a difference between model performance and it's accuracy? If yes, what exactly?

Secondly, what can I interpret from this classification_report of my model. Eg: The model's ability to predict 1 is 87% or 51%. Also, will accuracy be a good metric to evaluate as there's a major class imbalance but this class imbalance is of test data and not training, so I'm confused here as well?

enter image description here

I'm confused, is the model good a predicting 1 or 0

  1. you decide whether accuracy is a good metric to evaluate model performance based on your knowledge of your data and your aim (for example if your aim is to classify rare diseases accuracy is usually a bad metric, but that is not your case)
  2. if your model says an element is 1, then 87% of times he is right (not bad)
  3. where your model fails a bit is recall on 1, which means he only detected as 1 36% of all 1s available. so if your aim is to build a diagnostic test, where 1 means ill, you should improve the recall of your model
  4. f1-score is a quantity that tries to combine precision and recall in one number to penalize models that despite having high accuracy for example have low recall for one class https://en.wikipedia.org/wiki/Precision_and_recall

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.