INFORMATION & COMMUNICATION

Kullback–Leibler divergence

quantifies the information difference when describing P in terms of Q.

DKL(P∥Q)=∑xP(x)ln⁡P(x)Q(x)D_{KL}(P\Vert Q)=\sum_x P(x)\ln\frac{P(x)}{Q(x)}

symbols, variables and units

P,Q: discrete probability; D_KL: nat.

applicable conditions and boundaries

Where P(x)>0 requires Q(x)>0 to be finite; asymmetric, not a distance measure.

formula source code

The following is a copyable LaTeX expression.

D_{KL}(P\Vert Q)=\sum_x P(x)\ln\frac{P(x)}{Q(x)}

Reference and Extended Learning

MIT OpenCourseWare · Information Theory ↗

is organized according to model definition and assumptions. Please check actual conditions and original literature before engineering, research and clinical use.

KL divergencerelative entropy

Same subject formula

Go to Free Science Tool Library ↗