Accuracy

INFO

Measures the proportion of correct predictions out of total predictions.

How It Works

$$\text{Accuracy}=\frac{TP+TN}{TP+TN+FP+FN}$$

• TP: Correctly predicted positive cases
• TN: Correctly predicted negative cases
• FP: Incorrectly predicted positive cases
• FN: Incorrectly predicted negative cases

Accuracy reflects how often the model is correct across all predictions.

What to Look For

• Works well with balanced datasets
• Can be misleading with class imbalance
• Use alongside precision and recall for deeper insight

Application Models

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