Premium
Bayes two‐decision procedures for two‐sample problems and rank‐order data
Author(s) -
Baskerville J. C.,
Solomon D. L.
Publication year - 1975
Publication title -
canadian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.804
H-Index - 51
eISSN - 1708-945X
pISSN - 0319-5724
DOI - 10.2307/3315279
Subject(s) - bayes' theorem , rank (graph theory) , sample (material) , prior probability , mathematics , decision problem , set (abstract data type) , bayes factor , statistics , computer science , bayesian probability , algorithm , combinatorics , chromatography , chemistry , programming language
Abstract A two‐sample problem for rank‐order data is formulated as a two‐decision problem. Using the general Bayes solution, Bayes procedures are derived for several configurations of the set of states of nature including some for which the problem is distribution‐free. It is shown that for certain prior distributions these procedures reduce to classical LMP rank tests. Some devices for selection of prior distributions are suggested. It is shown that the Bayes risk of these procedures tends to zero as sample sizes increase.