This dissertation establishes a measure-theoretic foundation for information theory by formulating entropy, mutual information, and related quantities on a common type of probability spaces. Using Radon–Nikodym derivatives ...