Limit theorems for correlated Bernoulli random variables相关伯努利随机变量的极限定理.pdfVIP

Statistics and Probability Letters 78 (2008) 2339–2345 /locate/stapro Limit theorems for correlated Bernoulli random variables Barry James, Kang James, Yongcheng Qi ∗ Department of Mathematics and Statistics, University of Minnesota Duluth, 1117 University Drive, Duluth, MN 55812, USA Received 4 April 2006; received in revised form 8 January 2008; accepted 21 January 2008 Available online 26 February 2008 Abstract In this paper we study limit theorems for a class of correlated Bernoulli processes. We obtain the strong law of large numbers, central limit theorem and the law of the iterated logarithm for the partial sums of the Bernoulli random variables. c 2008 Elsevier B.V. All rights reserved. 1. Introduction Consider a Bernoulli process {X j , j ≥ 1} in which the random variables X j are correlated in the sense that the success probability of a trial conditional on the previous trials depends on the total number of successes achieved to that point. More precisely, assume that for some 0 p 1, P (X = 1|F ) = (1 − θ )p + θ j −1S , j +1 j j j j where 0 ≤ θj ≤ 1 are dependence parameters, Sn = n X j for n ≥ 1, and Fn = σ (X 1, . . . , X n ), the σ - j =1 field generated by X 1, . . . , X n . If X 1 has a Bernoulli distribution with parameter p , it follows that X 1, X 2 , . . . are identically distributed Bernoulli random variables. This process was introduced by Drezner and Farnum (1993), and the distribution of Sn is called a general