Summary
If the sum of two independent k-dimensional non-negative random variables is distributed as a multinomial, it is shown that the summands must be individually multinomial variates with the same set of probabilities.
Similar content being viewed by others
References
Cramer, H. (1936).—Uber eine Eigenschaft der normalen Verteilungsfunktion. Math. Zs., 41, 405–414.
Lukacs, E. (1970).—Characteristic functions, Griffin, London 2nd edition.
Raikov, D. (1937).—On the characterization of the Poisson laws, C. R. Acad. Sci. U.S.S.R., 14, pp. 8–11.
Rights and permissions
About this article
Cite this article
Shanbhag, D.N., Basawa, I.V. On a characterization property of the multinomial distribution. Trab. Estad. Invest. Oper. 25, 109–112 (1974). https://doi.org/10.1007/BF03030154
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF03030154