A characterization of entropy in terms of information loss (Q400959)

There are a slew of characterizations of Shannon entropy [\textit{C. E. Shannon}, Bell Syst. Tech. J. 27, 379--423, 623--656 (1948; Zbl 1154.94303)] and Tsallis entropy [\textit{C. Tsallis}, J. Stat. Phys. 52, No. 1--2, 479--487 (1988; Zbl 1082.82501)] as measures of information pervious to certain properties. Making use of the work by \textit{D. K. Faddeev} [Usp. Mat. Nauk 11, No. 1(67), 227--231 (1956; Zbl 0071.13103)] and \textit{S. Furuichi} [IEEE Trans. Inf. Theory 51, No. 10, 3638--3645 (2005; Zbl 1298.94038)], this paper derives a pretty simple characterization. The main result is that Shannon entropy is the only concept of information loss that is functorial, convex-linear and continuous, which is to be naturally generalized to Tsallis entropy as well.