Recently I talked about metrics for inbalanced binary classification problems.
I had trouble properly explaining the Precision-Recall Curve in simple words.
Sometimes a nice visualization tells more than a thousand words.
So, I made this simple interactive chart:
\(Precision=\frac{TP}{TP+FP}\)
\(Recall=\frac{TP}{TP+FN}\)
\(TPR=\frac{TP}{TP+FN}\)
\(FPR=\frac{FP}{TN+FP}\)
References
- Formulas: F-score
- Original notebook: Interactive Precision-Recall Curve