Queen's University - Utility Bar

QSpace at Queen's University >
Theses, Dissertations & Graduate Projects >
Queen's Theses & Dissertations >

Please use this identifier to cite or link to this item: http://hdl.handle.net/1974/492

Title: Nonparametric statistical procedures for therapeutic clinical trials with survival endpoints
Authors: Luo, Yingchun

Files in This Item:

File Description SizeFormat
Luo_Yingchun_200707_PhD.pdf1.18 MBAdobe PDFView/Open
Keywords: Nonparametric statistics
Equivalence clinical trial
Time to event endpoint
Issue Date: 2007
Series/Report no.: Canadian theses
Abstract: This thesis proposed two nonparametric statistical tests, based on the Kolmogorov-Smirnov distance and L2 mallows disatnce. To implement the proposed tests, nonparametric bootstrap method is employed to approximate the distributions of the test statistics to construct the corresponding bootstrap confidence interval procedures. Monte-Carlo simulations are performed to investigate the actual type I error of the proposed bootstrap procedures. It is found that the type I error of the bootstrap BC confidence interval procedure is close to the nominal level when censoring is not heavy and the boosttrap percentile confidence interval procedure works well when Kolmogorov-Smirnov distance is used to characterize the equivalence. When the data is heavily censored, the procedures based on the Kolmogorov-Smirnov distance have very conservative type I errors, while the procedures based on the Mallows distance are very liberal.
Description: Thesis (Ph.D, Mathematics & Statistics) -- Queen's University, 2007-08-01 10:43:32.345
URI: http://hdl.handle.net/1974/492
Appears in Collections:Queen's Theses & Dissertations
Mathematics & Statistics Graduate Theses

Items in QSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

 

  DSpace Software Copyright © 2002-2008  The DSpace Foundation - TOP