Summary
LetF be a distribution function over the real line. DefineR p(y)=∫|x−y|pdF(x) forp≧1. Forp>1 there is a unique minimizer ofR p(·), sayγ p. Under mild conditions onF it is shown that\(\mathop {\lim }\limits_{p \to 1} \gamma _p \) exists and that the limit is a median. Thus, one could use this limit as a definition of a unique median. Also it is shown that\(\mathop {\lim }\limits_{p \to 1} \gamma _p = {{\left( {R + L} \right)} \mathord{\left/ {\vphantom {{\left( {R + L} \right)} 2}} \right. \kern-\nulldelimiterspace} 2}\) whereR is the right extremity ofF andL is the left extremity ofF provided that −∞<L≦R<∞. A similar result is available ifL=−∞,R=∞, yetF has symmetric tails.
Similar content being viewed by others
References
DeGroot, M. H. and Rao, M. M. (1963). Bayes estimation with convex loss,Ann. Math. Statist.,34, 839–846.
Jackson, D. (1921). Note on the median of a set numbers,Bull. Amer. Math. Soc.,27, 160–164.
Author information
Authors and Affiliations
About this article
Cite this article
Umbach, D. A note on the median of a distribution. Ann Inst Stat Math 33, 135–140 (1981). https://doi.org/10.1007/BF02480927
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/BF02480927