BOOTSTRAPPED MULTI-DIMENSIONAL PRODUCT LIMIT PROCESS |
| |
| Authors: | Brian S Yandell Lajos Horváth |
| |
| Institution: | Departments of Statistics and of Horticulture, University of Wisconsin-Madison, Madison, Wisconsin, USA 53706;Department of Statistics, University of Wisconsin-Madison |
| |
| Abstract: | We approximate the limit process for a multivariate censored survival distribution using the bootstrap. The empirical process has a complicated covariance structure depending on the survival and censoring. The bootstrapped process provides a means to develop distribution-free procedures including simultaneous confidence bands. Results extend to comparison of multivariate survival distributions. A gallstone study is examined in some detail. |
| |
| Keywords: | Survival random censorship strong approximation cholecystectomy |
|
