The occurrence of sequence patterns in repeated experiments and hitting times in a Markov chain
暂无分享,去创建一个
A martingale argument is used to derive the generating function of the number of i.i.d. experiments it takes to observe a given string of outcomes for the first time. Then, a more general problem can be studied: How many trials does it take to observe a member of a finite set of strings for the first time? It is shown how the answer can be obtained within the framework of hitting times in a Markov chain. For these, a result of independent interest is derived.
[1] Ronald A. Howard,et al. Dynamic Probabilistic Systems , 1971 .
[2] Kai Lai Chung,et al. Markov Chains with Stationary Transition Probabilities , 1961 .
[3] Leonidas J. Guibas,et al. String Overlaps, Pattern Matching, and Nontransitive Games , 1981, J. Comb. Theory, Ser. A.
[4] Thinning of Renewal Point Processes: A Flow Graph Study. , 1974 .
[5] William Feller,et al. An Introduction to Probability Theory and Its Applications , 1967 .