再發事件資料常見於醫學、工業、財經、社會等等領域中，對再發資料分析研究時，我們往往無法確知再發事件發生的時間或是發生次數的分配。因此，本論文探討的是分析再發事件的無母數方法，包括Nelson提出的平均累積函數(mean cumulative function)估計量，及Wang、Chiang與Huang介紹的發生率(occurrence rate)之核函數(kernel function)估計量。
就平均累積函數估計量來說，藉由Nelson導出的變異數及自然(naive)變異數，可分別求得平均累積函數的區間估計。本文利用靴環法(bootstrap)計算出平均累積函數在不同時點的變異數，再與Nelson變異數及自然變異數比較，結果顯示Nelson變異數與靴環法算出的變異數較接近。因此，應依據Nelson變異數建構出事件發生累積次數之漸近信賴區間。
本論文亦介紹了兩個或多個母體的平均累積函數的比較方法，包含固定時點之比較與整條曲線之比較。在固定時點之下，比較方法分別為平均累積函數成對差異之漸近信賴區間及靴環信賴區間、變異數分析比較法，與排列檢定法；而整條曲線比較方法包含：類似 統計量、Lawless-Nadeau檢定。這些方法應用在本論文所採之實證資料時，所得到的檢定結論是一致的。

Recurrent event data arise in many fields, such as medicine, industry, economics, social sciences and so on. When studying recurrent event data, we usually don’t know the exact joint or marginal distributions of the occurrence times or the number of events over time. So, in this article we talk about some nonparametric methods, such as the mean cumulative function (MCF) discussed by Nelson, and kernel estimation of the rate function introduced by Wang, Chiang and Huang.
As to the estimator of MCF, we can compute the confidence interval by Nelson’s variance and naive variance. We use bootstrap method to compare the performance of Nelson variance of the estimated MCF and naive variance of the estimated MCF. The results show that Nelson variance is better than naive variance, so we should construct the confidence limits for the MCF by Nelson’s variance except when only grouped data are available.
We also introduce methods for comparing MCFs, including pointwise comparison of MCFs and comparison of entire MCFs. Methods for pointwise comparing MCFs include approximate confidence limits for difference between two MCFs, analysis-of-variance comparison, permutation test, and bootstrap’s confidence limits for difference between two MCFs. Methods for comparing entire MCFs include a statistic like Hoetelling’s , and Lawless-Nadeau test. Finally, all approaches are employed to analyze a real data, and the conclusions concordance with each other.

Aalen, O. O. (1978). Nonparametric inference for a family of counting processes, Annals of Mathematical Statistics, 6, 701-726.

Andersen, K. Borgan, O. Gill, R. D. Keiding, N. (1993). Statistical Models Based On Counting Processes, Springer-Verlag, New York.

Andrews, D. F. and Herzberg, A. M. (1985). A Collection of Problems from Many fields for The Student and Research Worker. Springer-Verlag, New York.

Ascher, H. and Feingold, H. (1984). Repairable Systems Reliability, Marcel Dekker, New York.

Bartoszyński, R., Brown, B. W., McBride, C. M. and Thompson, J. R. (1981). Some nonparametric techniques for estimating the intensity function of a cancer related nonstationary Poisson process. Annals of Statistics, 9 , 1050-1060.

Davis, D. J. (1952). An analysis of some failure data. Journal of The American Statistical Association, 47, 113-150.

Davison, A. C. and Hinkley, D. V. (1997). Bootstrap Methods and Their Application. Cambridge, UK.

Efron, B. (1979). Boostrap methods: another look at the jackknife. Annals of Statistics, 7, 1-26.

Engelhardt, M. (1995). Models and analyses for the reliability of a single repairable system. Recent Advances in Life-Testing and Reliability, CRC Press, Boca Raton, FL, 79-106.

Fleming, T. R. and Harrington, D. P. (1991). Counting Processes and Survival Analysis, John Wiley, New York.

Lancaster, T., Intrator, O. (1998). Panel Data With Survival: Hospitalization of HIV Patients. Journal of The American Statistical Association, 93, 46-53.

Lawless, J. F. (1987). Regression methods for poisson process data. Journal of The American Statistical Association, 82,808-815.

Lawless, J. F. and Nadeau C. (1995). Some simple Robust Methods for the analysis of recurrent events. Technometrics, 37,158-167.

Lin, D. Y., Wei, L. J. and Ying, Z. (2000). Semiparametric regression for the mean and rate functions of recurrent events. Journal of the Royal Statistical Society, Series B 62, 711-730.

Morrison, D. F. (1990). Multivariate Statistical Methods. Duxbury Advanced Series.

Nelson, W. (1988). Graphical analysis of system repair data, Journal of Quality Technology, 20, 24-35.

Nelson, W. (1995). Confidence limits for recurrence data: Applied to cost or number of repairs. Technometrics, 37,147-157.

Nelson, W.B. (2003). Recurrent events data analysis for product repairs, disease recurrences, and other applications. ASA-SIAM Series on Statistics and Applied Probability 10.

Pepe, M. S. and Cai, J. (1993). Some graphical displays and marginal regression analysis for recurrent failure times and time dependent covariates. Journal of the American Statistical Association. 88, 811-820.

Phillips, M. J. (2000). Bootstrap confidence regions for the expected ROCOF of a repairable system. IEEE Transactions on Reliability, 49,204-208.

Rigdon, S. E. and Basu, A. P. (2000). Statistical Methods for the Reliability of Repairable Systems, John Wiley, New York.

Sheather, S. J. and Jones, M. C. (1991). A reliable date-based bandwidth selection method for kernel density. Journal of the Royal Statistical Society, Series B, 53, 683-690.

Silverman, B. W. (1986). Density Estimation for Statistics and Data Analysis, Chapman and Hall, London.

Simonoff, J. S. (1996). Smoothing Methods in Statistics. Springer, New York.

Wang, M. C., Chiang, C. T. (2001). Analyzing recurrent event data with informative censorings. Journal of the American Statistical Association , 96,1057-1065.

Wang, M. C., Chiang, C. T. (2002). Non-parametric methods for recurrent event data with informative and non-informative censorings. Statistics in Medicine, 21,445-456.

Wang, M. C., Chiang, C. T. and Huang, C.Y. (2005). Kernel estimation of rate function for recurrent event data. Foundation of the Scandinavian Journal of Statistics, 32, 77-91.

Wang, M. C., Chiang, C. T., James, L. F. (2005). Random weighted bootstrap method for recurrent events with informative censoring. Lifetime Data Analysis, 11,489-509.