TY - JOUR

T1 - A characterization of conjugate priors in exponential families with application to inverse regression

AU - Luo, Wei

AU - Altman, Naomi S.

N1 - Funding Information:
The authors thank the editor and an anonymous reviewer for their proactive comments. The authors acknowledge partial support from NSF DMS 1007801 . Naomi Altman also acknowledges partial support from NSF DMS DMS-1127914 to the Statistical and Applied Mathematical Sciences Institute and from NIH UL1RR033184 .

PY - 2013/2

Y1 - 2013/2

N2 - It is often convenient to assume that X and X {pipe} Y are in the same exponential family. By considering X as the "parameter" and Y as the "data", the problem becomes determining which exponential families X {pipe} Y have conjugate priors. We develop a necessary condition for conjugacy. One-dimensional exponential families can be conjugate only if they have exactly two support points.

AB - It is often convenient to assume that X and X {pipe} Y are in the same exponential family. By considering X as the "parameter" and Y as the "data", the problem becomes determining which exponential families X {pipe} Y have conjugate priors. We develop a necessary condition for conjugacy. One-dimensional exponential families can be conjugate only if they have exactly two support points.

UR - http://www.scopus.com/inward/record.url?scp=84875121200&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84875121200&partnerID=8YFLogxK

U2 - 10.1016/j.spl.2012.10.030

DO - 10.1016/j.spl.2012.10.030

M3 - Article

AN - SCOPUS:84875121200

SN - 0167-7152

VL - 83

SP - 650

EP - 654

JO - Statistics and Probability Letters

JF - Statistics and Probability Letters

IS - 2

ER -