Index: The Book of Statistical ProofsGeneral TheoremsInformation theoryDiscrete mutual information ▷ Relation to joint and conditional entropy

Theorem: Let $X$ and $Y$ be discrete random variables with the joint probability $p(x,y)$ for $x \in \mathcal{X}$ and $y \in \mathcal{Y}$. Then, the mutual information of $X$ and $Y$ can be expressed as

\[\label{eq:dmi-jce} \mathrm{I}(X,Y) = \mathrm{H}(X,Y) - \mathrm{H}(X|Y) - \mathrm{H}(Y|X)\]

where $\mathrm{H}(X,Y)$ is the joint entropy of $X$ and $Y$ and $\mathrm{H}(X \vert Y)$ and $\mathrm{H}(Y \vert X)$ are the conditional entropies.

Proof: The existence of the joint probability mass function ensures that the mutual information is defined:

\[\label{eq:MI} \mathrm{I}(X,Y) = \sum_{x \in \mathcal{X}} \sum_{y \in \mathcal{Y}} p(x,y) \log \frac{p(x,y)}{p(x)\,p(y)} \; .\]

The relation of mutual information to conditional entropy is:

\[\label{eq:dmi-mce1} \mathrm{I}(X,Y) = \mathrm{H}(X) - \mathrm{H}(X|Y)\] \[\label{eq:dmi-mce2} \mathrm{I}(X,Y) = \mathrm{H}(Y) - \mathrm{H}(Y|X)\]

The relation of mutual information to joint entropy is:

\[\label{eq:dmi-mje} \mathrm{I}(X,Y) = \mathrm{H}(X) + \mathrm{H}(Y) - \mathrm{H}(X,Y) \; .\]

It is true that

\[\label{eq:MI-s1} \mathrm{I}(X,Y) = \mathrm{I}(X,Y) + \mathrm{I}(X,Y) - \mathrm{I}(X,Y) \; .\]

Plugging in \eqref{eq:dmi-mce1}, \eqref{eq:dmi-mce2} and \eqref{eq:dmi-mje} on the right-hand side, we have

\[\label{eq:MI-s2} \begin{split} \mathrm{I}(X,Y) &= \mathrm{H}(X) - \mathrm{H}(X|Y) + \mathrm{H}(Y) - \mathrm{H}(Y|X) - \mathrm{H}(X) - \mathrm{H}(Y) + \mathrm{H}(X,Y) \\ &= \mathrm{H}(X,Y) - \mathrm{H}(X|Y) - \mathrm{H}(Y|X) \end{split}\]

which proves the identity given above.

Sources:

Metadata: ID: P21 | shortcut: dmi-jce | author: JoramSoch | date: 2020-01-13, 22:17.