Language:   Search:   Contact
Zentralblatt MATH has released its new interface!
For an improved author identification, see the new author database of ZBMATH.

Query:
Fill in the form and click »Search«...
Format:
Display: entries per page entries
Zbl 1092.62107
Sun, Jianguo; Zhao, Qiang; Zhao, Xingqiu
Generalized log-rank tests for interval-censored failure time data.
(English)
[J] Scand. J. Stat. 32, No. 1, 49-57 (2005). ISSN 0303-6898; ISSN 1467-9469/e

A class of tests is considered for the $k$-samples homogeneity hypothesis by interval censored failure time data. I.e., for each subject a random interval $(L_i,R_i)$ is observed to which its failure time belongs. The test statistics is $$U_\xi=\sum_{i=1}^n x_i{ \xi(\widehat G_n(L_i))-\xi(\widehat G_n(R_i)) \over \widehat G_n(L_i)-\widehat G_n(R_i)},$$ where $n$ is the number of subjects in the union of all samples, $x_i$ is the vector of indicators of the sample (its $l$-th element equals 1 iff the $i$-th subject belongs to the $l$-th sample and is 0 otherwise), $\widehat G_n(x)$ is a nonparametric estimator of the survival function under the null hypothesis (homogeneity), $\xi$ is a fixed function. (E.g., for $\xi(x)=x\log(x)$ this is the score statistics). The asymptotic normality of $U_\xi$ under $H_0$ is demonstrated. Simulation results and real breast cancer data application are considered.
[R. E. Maiboroda (Ky\"iv)]
MSC 2000:
*62N03 Testing
62G10 Nonparametric hypothesis testing
62G20 Nonparametric asymptotic efficiency

Keywords: multisample homogeneity test; survival function; score test; asymptotic normality

Highlights
Master Server