⚖️ Precision-Recall Tradeoff

Understand the balance between precision and recall in classification

Classification Threshold: 0.50

Class Separation: 1.5

Score Distribution

Precision-Recall Curve

0%

Precision

0%

Recall (TPR)

0%

F1 Score

0

True Positives

0

False Positives

0

False Negatives

📚 Understanding Precision-Recall Tradeoff

Precision: Of predicted positives, how many are correct? TP / (TP + FP)
Recall: Of actual positives, how many detected? TP / (TP + FN)
Tradeoff: Higher threshold → higher precision, lower recall
F1 Score: Harmonic mean balancing both
Use Case: Spam detection (high precision) vs Cancer screening (high recall)