I'm trying to evaluate a binary classifier. I'm not sure if it makes sense to calculate the f1-measure for each class separately or if it would be better to calculate the overall f1-measure.

Can you help me?

CodePudding user response：

f1-measure balances accuracy and recall. Depending on the problem and data, you can compute the f1-measure for each class or the total measure.

If one class has many more instances than the other, calculating the f1-measure for each class might be more revealing. This shows how well the classifier performs on the minority class, which is frequently the class of interest.

If the class distribution is balanced or you're interested in the classifier's overall performance, compute the overall f1-measure.

Always consider precision, recall, accuracy, and ROC curve.