設{ Xn ; n ? o } 為一不可約，單一週期及正常返(irreducible, aperiodicand positive recurrent) 的Markov 鏈，令其狀態空間(state space) S 為有限及移轉矩陣(transitionprobability matrix) P. 假設Pij’S 是某一個未知參數θ的函數，且此函數為已知，稱之為PjJ (θ) ，其值域屬於RK ，並令π是它的平接機率(stationary probability) ，給定一組己知的觀察值{ Xn; 0 ?j? n }，令θn 為θ 的最大概似估計(maxirnumlikelihood estimator). 在本論文中，我們要利用自助法

(bootstrap method) 去逼近θn 的樣本分佈，並證明√n(θn-θ)的自助樣本分佈(bootstrap sampling distribution) 和其本身之樣本分佈是漸近相合的(asyrnptotically closed). 最後，並以並以模擬的數值作為參考.

References

1. Anderson, T. and Goodman, L. (1957). Statistical inference about Markov chains.
Ann. Math. Statist. 28, pp. 89-110.
2. Athreya, K. B. and Fuh, C. D. (1989). Bootstrapping Markov chains: Countable
case. Technical Report: B-89-7, Institute of Statistical Science, Academia Sincia,
Taipei, Taiwan, ROC.
3. Bartlett, M. S. (1951). The frequency goodness of fit test for probability chains. Proc.Carnb. Phil. Soc. 47, pp. 86-95.
4. Basawa, I. V. and Prakasa Rao, B. L. S. (1980). Statistical inference for stochastic
processes. Academic Press, New York.
5. Billingsley, P. (1961). Statistical inference for Markov Processes. The University of
Chicago Press, Chicago.
6. Cinlar, E. (1975). Introduction to Stochastic Processes. pp. 106-168. Prentice-Hall,
Englewwood Cliffs, N. J.
7. Doob, J. L. (1953). Stochastic processes. p.228 and p.232. Springer- Verlag, Berlin.
8. Efron, B. (1982). The Jackknife, the Bootstrap and Other Resampling Plans. SIAM,
Philadelphia.
9. Kulperger, R. J. and Prakasa Rao, B. L. S. (1990). Bootstrapping a finite state
Markov chains. Sankhya A 51, Pt. 2, pp. 178-191.
10. Kemeny, J. G. and Snell, J. L. (1960). Finite Markov chains. pp. 69-71. SpringerVerlage,New York.
11. Rao, C. R. (1973). Linear statistkal inference and its applications. pp. 363-366.
Wieley: New York.