Title :
Asymptotic Geometry of Multiple Hypothesis Testing
Author :
Westover, M. Brandon
Author_Institution :
Dept. of Neurology, Massachusetts Gen. Hosp., Boston, MA
fDate :
7/1/2008 12:00:00 AM
Abstract :
We present a simple geometrical interpretation for the solution to the multiple hypothesis testing problem in the asymptotic limit. Under this interpretation, the optimal decision rule is a nearest neighbor classifier on the probability simplex.
Keywords :
error statistics; pattern recognition; probability; asymptotic limit; geometrical interpretation; multiple hypothesis testing; nearest neighbor classifier; optimal decision rule; probability simplex; Error probability; Histograms; Hospitals; Information geometry; Nearest neighbor searches; Nervous system; Pattern recognition; Probability distribution; Statistical learning; Testing; Geometry; hypothesis testing; large deviations; pattern recognition;
Journal_Title :
Information Theory, IEEE Transactions on
DOI :
10.1109/TIT.2008.924656
