Entropy and Mutual
Information
Let:
-
X be a random variable (R.V).
-
P(X) be the probability distribution of X.
-
p(x) be the probability density of X.
The entropy of X, H(X) is defined by:
H(X) = -EX[ log(P(X)) ]
Entropy is a measure of randomness. The more random a variable is, the more entropy it
will have.
Example:
The joint entropy is a statistics that summarizes the degree of dependence of a RV X
on an other RV Y. It is defined by:
H(X,Y) = -EX [EY [ log(P(X,Y)) ] ]
The conditional entropy is a statistics that summarizes the
randomness of Y given knowledge of X. It is defined by:
H(Y|X) = -EX [EY [ log(P(Y|X)) ] ]
Two random variables are considered to be independent if:
H(X,Y) = H(X) + H(Y)
The Mutual Information, MI, between two random variables X and Y is given by:
MI(X,Y) = H(Y) - H(Y|X) = H(X) + H(Y) - H(X,Y)
(It's thus a measure of the reduction of the entropy of Y given X.)