Abstract
A version of the multiple decsion problem is studied in which the procedure is based only on the current observation and the previous decision. A necessary and sufficient condition for inconsistency of the stepwise maximum likelihood procedure is shown to be the boundedness of the likelihood ratios. In the case of consistency the (typically slow) rate of convergence to zero of the error probabilities is determined.
Similar content being viewed by others
References
Cover T M (1969) Hypothesis testing with finite statistics.Ann. Math. Statist. 40: 828–835.
Cover T M, Freedman M A and Hellman M E (1976) Optimal finite memory learning algorithms for the finite sample problem.Inf Control 30: 49–85.
Hellman M E, Cover T M (1970) Learning with finite memory.Ann Math Statist 41: 765–782.
Nevelson M B, Hasminski R Z (1971)Stochastic Approximation and Recursive Estimation. American Mathematical Society. Providence, RI.
Renyi A (1969) On some problems of statistics from the point of view of information theory. inProceedings of the Colloquium on Information Theory, Debrecen, 343–357.
Rukhin A L (1993) Lower bound on the error probability for families with bounded likelihood ratios.Proc Amer Math Soc 119: 1307–1314.
Rukhin A L (1994) Recursive testing of multiple hypotheses: consistency and efficiency of the Bayes rule.Ann. Statist 22
Yakowitz S (1974) Multiple hypothesis testing by finite memory algorithms.Ann. Statist. 2: 323–336.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Rukhin, A.L., Shi, J. Recursive procedures for multiple decisions: finite time memory and stepwise maximum likelihood procedure. Stat Papers 36, 155–162 (1995). https://doi.org/10.1007/BF02926028
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/BF02926028