Recall

INFO

Measures the proportion of correctly predicted positive observations to all actual positives.

How It Works

$$\text{Recall}=\frac{TP}{TP+FN}$$

• TP: Correctly predicted positive cases
• FN: Missed actual positive cases

Recall measures how well the model captures actual positives.

What to Look For

• High recall = fewer false negatives
• Crucial in medical, fraud, or safety-critical domains
• Often traded off against precision

Application Models

• Convolutional Neural Network (CNN)
• Deep Neural Network (DNN)
• Transformer