TREE-STRUCTURED EXTREME-VALUE MODEL REGRESSION

被引：3

作者：

AHN, HS ^{[1
]}

机构：

[1] NATL CTR TOXICOL RES,DIV BIOMETRY & RISK ASSESSMENT,JEFFERSON,AR 72079

来源：

COMMUNICATIONS IN STATISTICS-THEORY AND METHODS | 1994年 / 23卷 / 01期

基金：

美国国家科学基金会;

关键词：

BOOTSTRAP; RECURSIVE PARTITIONING; SURVIVAL ANALYSIS; EXPONENTIAL REGRESSION; WEIBULL REGRESSION;

D O I：

10.1080/03610929408831245

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

This paper gives a method for fitting piecewise Weibull regression models to censored survival data. Recursive stratification is performed using statistical tests for variances and residual analysis. The average squared error of the residuals is used as a splitting criterion in cross-validation. The bootstrap resampling is used to keep the probability of a type I error of the method (the error of finding two or more strata when there is one). Real and simulated data are used for illustration.

引用

页码：153 / 174

页数：22

共 15 条

[1]

AHN H, IN PRESS BIOMETRICS

[2]

AHN H, 1992, THESIS U WISCONSINM

[3]

BUCKLEY J, 1979, BIOMETRIKA, V66, P429

[4]

COX DR, 1972, J R STAT SOC B, V34, P187

[5]

DICKSON ER, 1989, NEW ENGL J MED, V10, P1

[6]

Fleming T. R., 1991, COUNTING PROCESSES S

[7]

Levene, 1961, CONTRIBUTIONS PROBAB, P279, DOI DOI 10.1137/1003016

[8]

LIN DY, 1992, 111 U WASH DEPT BIOS

[9] SURVIVAL MODELING THROUGH RECURSIVE STRATIFICATION [J].

LOH, WY .

COMPUTATIONAL STATISTICS & DATA ANALYSIS, 1991, 12 (03) :295-313

[10]

MILLER R, 1982, BIOMETRIKA, V69, P521

← 1 2 →