Journal of The Japanese Society for Quality Control
Online ISSN : 2432-1044
Print ISSN : 0386-8230
Search
Advanced search
OR
Browse
Search
Volume 12, Issue 1
Displaying 1-13 of 13 articles from this issue
Show all abstracts
    • |<
    • <
    • 1
    • >
    • >|
Message
Article
Lecture
Contributed Paper
  • Statistical Analysis of Incomplete Data in Competing Risks Model
    Masami MlYAKAWA
    Article type: Contributed Paper
    1982 Volume 12 Issue 1 Pages 23-29
    Published: January 15, 1982
    Released on J-STAGE: March 07, 2019
    DOIhttps://doi.org/10.20684/quality.12.1_23
    JOURNAL RESTRICTED ACCESS
    Show abstractHide abstract
    In the statistical analysis of the field life data, we must sometimes estimate the failure rate of each component of a system by the system life data. In this paper a competing risks model has been used for determining the lifetimes of the components and two types of incomplete life data are considered. Assuming that the lifetimes of the components are independent and exponentially distributed variates, the maximum likelihood estimators or the uniformly minimum variance unbiased estimators for the failure rate of a particular component are derived, and the efficiencies of these estimators are also discussed. The analysis shows that when the ratio of the failure rate of the particular component and that of the other component, say ρ=λ/μ, is large, the reduction in efficiencies of estimators obtained by using the incomplete system life data is not so much as compared to that obtained by using the complete data.
    View full abstract
    Download PDF (623K)
  • On Inspection with Possible Overlooking of Defectives
    Yoritake Fujino
    Article type: Contributed Paper
    1982 Volume 12 Issue 1 Pages 31-40
    Published: January 15, 1982
    Released on J-STAGE: March 07, 2019
    DOIhttps://doi.org/10.20684/quality.12.1_31
    JOURNAL RESTRICTED ACCESS
    Show abstractHide abstract
    In the inspection of a lot of products, it occurs rather often that all of the defective items can not be detected for some reason or other. When we re-inspect the items which were once looked on as non-defective, both the probability of overlooking the defective items and the number of defectives in the lot are estimable. In this case the method of maximum likelihood and that of conditional maximum likelihood are shown to be effective if the probability of overlooking is not so large. The effects of variation of the probability upon the estimators are also investigated.
    View full abstract
    Download PDF (816K)
Activity Report of Research Section
Material
Case Method
    • |<
    • <
    • 1
    • >
    • >|
feedback
 Top