Data Science & ML
Advanced

Binary Cross-Entropy Loss

Loss function for binary classification.

Formula

L=−1n∑[ylog⁡(y^)+(1−y)log⁡(1−y^)]L = -\frac{1}{n}\sum[y\log(\hat{y}) + (1-y)\log(1-\hat{y})]

Variables

yTrue label (0 or 1)
\hat{y}Predicted probability
nSamples

Example

y=1, p=0.9: loss = -log(0.9) = 0.105

Did You Know?

Cross-entropy comes from information theory and measures surprise.

lossclassification