Premium
An Application of Generalized Linear Models to Survival Analysis with Two Types of Failure
Author(s) -
Hurley M. A.
Publication year - 1985
Publication title -
journal of the royal statistical society: series c (applied statistics)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.205
H-Index - 72
eISSN - 1467-9876
pISSN - 0035-9254
DOI - 10.2307/2347473
Subject(s) - mathematics , generalized linear model , statistics , econometrics
SUMMARY A theoretical framework is presented for the analysis of survival data where two types of failure occur and in which the number of failures of the first type is counted for each of a sequence of n time intervals but the number of failures of the second type is determined only for the whole period. Rates of failure of the first type are related to explanatory variables and experimental factors whereas those of the second type are considered as nuisance parameters. Two models are described, based on an approximate and the exact approach, and are applied to a Latin square experiment. The approximate model performs well compared to the exact model, provided that a 2 ‐n proportion of experimental units survive to the end of the experiment.