Test regions using two or more correlated product characteristics

Willem Albers, Gerda R.J. Arts, Wilbert C.M. Kallenberg

    Research output: Contribution to journalArticleAcademicpeer-review

    3 Citations (Scopus)
    33 Downloads (Pure)

    Abstract

    For inspection of manufactured parts, it may be much easier and/or cheaper to use the information from two or more product characteristics which are strongly related to the characteristic of interest, than to measure this characteristic itself. To get high quality output, it is required that from the accepted parts only a given, typically very small, fraction is nonconforming. In this paper test regions are determined such that, under the preceding condition, the number of accepted products is maximized. The good performance of the methods is illustrated by Monte Carlo results. An application in semiconductor industry shows its value in practice.
    Original languageEnglish
    Pages (from-to)153-165
    Number of pages9
    JournalTechnometrics
    Volume41
    Issue number2
    DOIs
    Publication statusPublished - 1999

    Keywords

    • Inspection
    • Simulations
    • Yield
    • Consumer risk

    Fingerprint

    Dive into the research topics of 'Test regions using two or more correlated product characteristics'. Together they form a unique fingerprint.

    Cite this