Data-transformation approach to lifetimes data analysis: An overview

• Published in: Journal of statistical planning and inference Vol. 140; no. 10; pp. 2904 - 2917
• Main Authors: Mudholkar, Govind S., Asubonteng, Kobby O.
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Kidlington Elsevier B.V 01.10.2010; Elsevier
• Subjects: Applications; Duration dependence; Exact sciences and technology; Exponentiated Weibull; General topics; Generalized Weibull; Mathematics; Maximum likelihood estimation; Multivariate analysis; Probability and statistics; Probability theory and stochastic processes; Reliability; Reliability, life testing, quality control; Sciences and techniques of general use; Special processes (renewal theory, markov renewal processes, semi-markov processes, statistical mechanics type models, applications); Statistics; Survival analysis; Transformations; Weibull distribution; Maximum likelihood estimation; Survival analysis; 62P20; Exponentiated Weibull; 62N02; Duration dependence; Generalized Weibull; 62N05; Transformations; Weibull distribution; Reliability; Statistical distribution; Data transformation; Prediction theory; Stochastic process; Economic sciences; Sampling distribution; Industry; Engineering; Approximation theory; Data analysis; Social sciences; Goodness of fit test; Finance; Statistical estimation; Statistical decision; Life test; Sociometry; Statistical method; Lifetime; Filtering theory; Maximum likelihood; Econometrics
• ISSN: 0378-3758; 1873-1171
• DOI: 10.1016/j.jspi.2010.03.014

This article consists of a review and some remarks on the scope, common models, methods, their limitations and implications for the analysis of lifetime data. Also a new approach based upon data-transformations analogous to that of Box and Cox (1964) is introduced. The basic methods and theory of the subject are most familiarly and commonly encountered by the statistical community in the context of problems in reliability studies and survival analysis. However, they are also useful in areas of statistical applications such as goodness-of-fit and approximations for sampling distributions and are applicable in such diverse fields of applied research as economics, finance, sociology, meteorology and hydrology. The discussion includes examples from the mainstream statistical, social sciences and business literature.