F1-Score (EN)

Concept

Metric for evaluating classification models that harmonically averages precision and recall

F1-Score Calculation

flowchart TD     A[True Positive TP] --> D[F1-Score]     B[False Positive FP] --> E[Precision
TP/(TP+FP)]     C[False Negative FN] --> F[Recall
TP/(TP+FN)]     E --> D     F --> D     D --> G[Harmonic Mean
2 × (Precision × Recall)/(Precision + Recall)]

In Context

Typically used together with Precision, Recall, and ROC-AUC
Related to: Confusion Matrix, Accuracy, Matthews Correlation Coefficient
Example use: Evaluation of spam filters, medical diagnosis models

Quelle: AI Generated