So berechnen Sie Confusion Matrix
learn.whatIsHeading
Creates confusion matrix showing actual vs. predicted classifications. Basis for evaluation metrics.
Formel
Accuracy = (TP+TN) / total
- TP
- TN/(TN+FP) — TN/(TN+FP)
- TN
- TN value — Variable used in the calculation
Schritt-für-Schritt-Anleitung
- 14 cells: TP (correct positive), FP (false positive), FN (false negative), TN (correct negative)
- 2Accuracy = (TP+TN) / total
- 3Sensitivity/Recall = TP/(TP+FN), Specificity = TN/(TN+FP)
- 4Precision = TP/(TP+FP)
Gelöste Beispiele
Eingabe
TP/FP/TN/FN
Ergebnis
Metrics calc
Häufige Fehler vermeiden
- ✕Using accuracy for imbalanced data (wrong)
- ✕Confusing sensitivity and specificity
- ✕Not balancing precision/recall tradeoff
Häufig gestellte Fragen
When use different metrics?
Accuracy: balanced classes; precision: minimize false positives; recall: minimize false negatives.
What about imbalanced classes?
Accuracy misleading; use precision, recall, F1-score, or AUC instead.
Bereit zur Berechnung? Probieren Sie den kostenlosen Confusion Matrix-Rechner aus
Probieren Sie es selbst aus →