Abstract
Let {X i āi=1 } be a sequence of independent Bernoulli random variables with probability p that X i =1 and probability q=1 ā p that X i =0 for all iā„1. We consider time-invariant finite-memory (i.e., finite-state) estimation procedures for the parameter p which take X 1, ... as an input sequence. In particular, we describe an n-state deterministic estimation procedure that can estimate p with mean-square error O(log n/n) and an n-state probabilistic estimation procedure that can estimate p with mean-square error O(1/n). We prove that the O(1/n) bound is optimal to within a constant factor. In addition, we show that linear estimation procedures are just as powerful (up to the measure of mean-square error) as arbitrary estimation procedures. The proofs are based on the Markov Chain Tree Theorem.
This research was supported by the Bantrell Foundation and by NSF grant MCS-8006938.
Preview
Unable to display preview. Download preview PDF.
References
B. Chandrasekaran and C. Lam, āA Finite-Memory Deterministic Algorithm for the Symmetric Hypothesis Testing Problem,ā IEEE Transactions on Information Theory, Vol. IT-21, No. 1, January 1975, pp. 40ā44.
T. Cover, āHypothesis Testing with Finite Statisticsā, Annals of Mathematical Statistics, Vol. 40, 1969, pp. 828ā835.
T. Cover and M. Hellman, āThe Two-Armed Bandit Problem With Time-Invariant Finite Memoryā, IEEE Transactions on Information Theory, Vol. IT-16, No. 2, March 1970, pp. 185ā195.
J. Doob, Stochastic Processes, Wiley, New York, 1953.
W. Feller, An Introduction to Probability Theory and its Applications, Wiley, New York, 1957.
P. Flajolet, āOn Approximate Counting,ā INRIA Research Report No. 153, July 1982.
R. Flower and M. Hellman, āHypothesis Testing With Finite Memory in Finite Time,ā IEEE Transactions on Information Theory, May 1972, pp. 429ā431.
M. Hellman, āFinite-Memory Algorithms for Estimating the Mean of a Gaussian Distribution,ā IEEE Transactions on Information Theory, Vol. IT-20, May 1974, pp. 382ā384.
M. Hellman and T. Cover, āLearning with Finite Memory,ā Annals of Mathematical Statistics, Vol. 41, 1970, pp. 765ā782.
W. Hoeffding and G. Simons, "Unbiased Coin Tossing with a Biased Coin,ā Annals of Mathematical Statistics, Vol. 41, 1970, pp. 341ā352.
J. Koplowitz, āNecessary and Sufficient Memory Size for m-Hypothesis Testing,ā IEEE Transactions on Information Theory, Vol. IT-21, No. 1, January 1975, pp. 44ā46.
J. Koplowitz and R. Roberts, āSequential Estimation With a Finite Statistic,ā IEEE Transactions on Information Theory, Vol. IT-19, No. 5, September 1973, pp. 631ā635.
S. Lakshmivarahan, Learning Algorithms ā Theory and Applications, Springer-Verlag, New York, 1981.
F. Leighton and R. Rivest, āThe Markov Chain Tree Theorem,ā to appear.
F. Morris, āCounting Large Numbers of Events in Small Registers,ā Communications of the ACM, Vol. 21, No. 10, October 1978, pp. 840ā842.
C. Mullis and R. Roberts, āFinite-Memory Problems and Algorithms,ā IEEE Transactions on Information Theory, Vol. IT-20, No. 4, July 1974, pp. 440ā455.
K. Narendra and M. Thathachar, āLearning Automata ā A Survey,ā IEEE Transactions on Systems, Vol. SMC-4, No. 4, July 1974, pp. 323ā334.
J. von Neumann, āVarious Techniques Used in Connection With Random Digits,ā Monte Carlo Methods, Applied Mathematics Series, No. 12, U.S. National Bureau of Standards, Washington D.C., 1951, pp. 36ā38.
A. Oppenheim and R. Schafe, Digital Signal Processing, Prentice-Hall, Englewood Cliffs, New Jersey, 1975.
L. Rabiner and B. Gold, Theory and Application of Digital Signal Processing, Prentice-Hall, Englewood Cliffs, New Jersey, 1975.
R. Roberts and J. Tooley, āEstimation With Finite Memory,ā IEEE Transactions on Information Theory, Vol. IT-16, 1970, pp. 685ā691.
A. Sage and J. Melsa, Estimation Theory With Applications to Communications and Control, McGraw-Hill, New York, 1971.
F. Samaniego, āEstimating a Binomial Parameter With Finite Memory,ā IEEE Transactions on Information Theory, Vol. IT-19, No. 5, September 1973, pp. 636ā643.
F. Samaniego, āOn Tests With Finite Menory in Finite Time,ā IEEE Transactions on Information Theory, Vol. IT-20, May 1974, pp. 387ā388.
F. Samaniego, āOn Testing Simple Hypothesis in Finite Time With Hellman-Cover Automata,ā IEEE Transactions on Information Theory, Vol. IT-21, No. 2, March 1975, pp. 157ā162.
B. Shubert, āFinite-Memory Classification of Bernoulli Sequences Using Reference Samples,ā IEEE Transactions on Information Theory, Vol. IT-20, May 1974, pp. 384ā387.
B. Shubert and C. Anderson, āTesting a Simple Symmetric Hypothesis by a Finite-Memory Deterministic Algorithm,ā IEEE Transactions on Information Theory, Vol. IT-19, No. 5, September 1973, pp. 644ā647.
T. Wagner, āEstimation of the Mean With Time-Varying Finite Memory,ā IEEE Transactions on Information Theory, Vol. IT-18, July 1972, pp. 523ā525.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
Ā© 1983 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Abstract, E. (1983). Estimating a probability using finite memory. In: Karpinski, M. (eds) Foundations of Computation Theory. FCT 1983. Lecture Notes in Computer Science, vol 158. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-12689-9_109
Download citation
DOI: https://doi.org/10.1007/3-540-12689-9_109
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-12689-8
Online ISBN: 978-3-540-38682-7
eBook Packages: Springer Book Archive