A family of test statistics for detecting monotone mean residual life

Bergman, Bo (author): Department of Mathematics, Linköping Institute of Technology

(creator_code:org_t)

Elsevier BV, 1989
1989
English.
In: Journal of Statistical Planning and Inference. - : Elsevier BV. - 0378-3758 .- 1873-1171. ; 21:2, s. 161-178

Related links:: https://urn.kb.se/re...; show more...; https://doi.org/10.1...; show less...

Abstract Subject headings

The mean residual life of a life distribution F with survival function F̄=1-F is defined as e(x)=(sh{phonetic}∞xF̄(t)dt)/F̄(x) if F̄(x)>0 (e(x)=0 if F̄(x)=0). The function e(x) has proved to be of great interest in many different areas such as biometry, reliability and social sciences. The class DMRL (IMRL) consists of those life distributions which have decreasing (increasing) mean residual life; i.e. e(x) is a decreasing (increasing) function of x. In this paper we present and study a family of test statistics intended for testing exponentiality against DMRL (IMRL). The family includes the test statistics proposed for the same purpose by Hollander and Proschan (1975) for complete samples and later generalized by Chen et al. (1983a) to accomodate randomly right censored samples. We compare the efficiency for different members in the family of test statistics when testing against some different alternatives. Furthermore, we study the efficiency loss due to different amount of exponential censoring. The asymptotical normal distribution of the test statistics are presented and the quality of the normal approximation is also discussed

About the subject

ENGINEERING AND TECHNOLOGY: ENGINEERING AND ...; and Mechanical Engin ...; and Reliability and ...

Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.