JensenShannon: Jensen-Shannon Divergence

In patrickreidy/distdist: Distances on Distributions

Description

Compute Jensen-Shannon divergences of two probability distributions.

Usage

## S4 method for signature 'Distribution,Distribution'
JensenShannon(p1, p2, a)

Arguments

p1, p2	Distributions.
a	A numeric constant between 0 and 1.

Details

Let p1 and p2 denote the vectors of probability mass assigned by two distributions defined on the same state space; furthermore, let these distributions be strictly positively-valued. Let p3 = a*p1 + (1-a)*p2, for a constant 0 < a < 1. Then, the Jensen-Shannon divergence is given by a*KullbackLeibler(p1, p3) + (1-a)*KullbackLeibler(p2, p3).

Jensen-Shannon divergence is a symmetric function only if a = 0.5.

See Also

Kullback-Leibler, Topsoe.

patrickreidy/distdist documentation built on May 22, 2019, 12:40 p.m.