A Robust Metric for Screening Outliers from Analogue Product Manufacturing Tests Responses

Shaji Krishnan, Shaji Krishnan, Hans G. Kerkhoff

    Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

    4 Citations (Scopus)

    Abstract

    Mahalanobis distance is one of the commonly used multivariate metrics for finely segregating defective devices from non-defective ones. An associated problem with this approach is the estimation of a robust mean and a covariance matrix. In the absence of such robust estimates, especially in the presence of outliers to test-response measurements, and only a sub-sample from the population is available, the distance metric becomes unreliable. To circumvent this problem, multiple Mahalanobis distances are calculated from selected sets of test-response measurements. They are then suitably formulated to derive a metric that has a reduced variance and robust to shifts or deviations in measurements. In this paper, such a formulation is proposed to qualitatively screen product outliers and quantitatively measure the reliability of the non-defective ones. The application of method is exemplified over a test set of an industrial automobile product.
    Original languageUndefined
    Title of host publicationProceedings 16th IEEE European Test Symposium, ETS 2011
    Place of PublicationUSA
    PublisherIEEE Computer Society
    Pages159-164
    Number of pages6
    ISBN (Print)978-1-4577-0483-3
    DOIs
    Publication statusPublished - 23 May 2011
    Event16th IEEE European Test Symposium, ETS 2011 - Trondheim, Norway
    Duration: 23 May 201127 May 2011
    Conference number: 16

    Publication series

    Name
    PublisherIEEE Computer Society
    ISSN (Print)1530-1877

    Conference

    Conference16th IEEE European Test Symposium, ETS 2011
    Abbreviated titleETS
    CountryNorway
    CityTrondheim
    Period23/05/1127/05/11

    Keywords

    • METIS-284958
    • IR-79215
    • manufacturing test
    • EWI-21152
    • Metrics
    • CAES-TDT: Testable Design and Test
    • Outliers

    Cite this

    Krishnan, S., Krishnan, S., & Kerkhoff, H. G. (2011). A Robust Metric for Screening Outliers from Analogue Product Manufacturing Tests Responses. In Proceedings 16th IEEE European Test Symposium, ETS 2011 (pp. 159-164). USA: IEEE Computer Society. https://doi.org/10.1109/ETS.2011.31
    Krishnan, Shaji ; Krishnan, Shaji ; Kerkhoff, Hans G. / A Robust Metric for Screening Outliers from Analogue Product Manufacturing Tests Responses. Proceedings 16th IEEE European Test Symposium, ETS 2011. USA : IEEE Computer Society, 2011. pp. 159-164
    @inproceedings{7051d36f49c04283a7e4681b86262c3a,
    title = "A Robust Metric for Screening Outliers from Analogue Product Manufacturing Tests Responses",
    abstract = "Mahalanobis distance is one of the commonly used multivariate metrics for finely segregating defective devices from non-defective ones. An associated problem with this approach is the estimation of a robust mean and a covariance matrix. In the absence of such robust estimates, especially in the presence of outliers to test-response measurements, and only a sub-sample from the population is available, the distance metric becomes unreliable. To circumvent this problem, multiple Mahalanobis distances are calculated from selected sets of test-response measurements. They are then suitably formulated to derive a metric that has a reduced variance and robust to shifts or deviations in measurements. In this paper, such a formulation is proposed to qualitatively screen product outliers and quantitatively measure the reliability of the non-defective ones. The application of method is exemplified over a test set of an industrial automobile product.",
    keywords = "METIS-284958, IR-79215, manufacturing test, EWI-21152, Metrics, CAES-TDT: Testable Design and Test, Outliers",
    author = "Shaji Krishnan and Shaji Krishnan and Kerkhoff, {Hans G.}",
    note = "Best paper award ETS 2011 ISSN: 1530-1877",
    year = "2011",
    month = "5",
    day = "23",
    doi = "10.1109/ETS.2011.31",
    language = "Undefined",
    isbn = "978-1-4577-0483-3",
    publisher = "IEEE Computer Society",
    pages = "159--164",
    booktitle = "Proceedings 16th IEEE European Test Symposium, ETS 2011",
    address = "United States",

    }

    Krishnan, S, Krishnan, S & Kerkhoff, HG 2011, A Robust Metric for Screening Outliers from Analogue Product Manufacturing Tests Responses. in Proceedings 16th IEEE European Test Symposium, ETS 2011. IEEE Computer Society, USA, pp. 159-164, 16th IEEE European Test Symposium, ETS 2011, Trondheim, Norway, 23/05/11. https://doi.org/10.1109/ETS.2011.31

    A Robust Metric for Screening Outliers from Analogue Product Manufacturing Tests Responses. / Krishnan, Shaji; Krishnan, Shaji; Kerkhoff, Hans G.

    Proceedings 16th IEEE European Test Symposium, ETS 2011. USA : IEEE Computer Society, 2011. p. 159-164.

    Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

    TY - GEN

    T1 - A Robust Metric for Screening Outliers from Analogue Product Manufacturing Tests Responses

    AU - Krishnan, Shaji

    AU - Krishnan, Shaji

    AU - Kerkhoff, Hans G.

    N1 - Best paper award ETS 2011 ISSN: 1530-1877

    PY - 2011/5/23

    Y1 - 2011/5/23

    N2 - Mahalanobis distance is one of the commonly used multivariate metrics for finely segregating defective devices from non-defective ones. An associated problem with this approach is the estimation of a robust mean and a covariance matrix. In the absence of such robust estimates, especially in the presence of outliers to test-response measurements, and only a sub-sample from the population is available, the distance metric becomes unreliable. To circumvent this problem, multiple Mahalanobis distances are calculated from selected sets of test-response measurements. They are then suitably formulated to derive a metric that has a reduced variance and robust to shifts or deviations in measurements. In this paper, such a formulation is proposed to qualitatively screen product outliers and quantitatively measure the reliability of the non-defective ones. The application of method is exemplified over a test set of an industrial automobile product.

    AB - Mahalanobis distance is one of the commonly used multivariate metrics for finely segregating defective devices from non-defective ones. An associated problem with this approach is the estimation of a robust mean and a covariance matrix. In the absence of such robust estimates, especially in the presence of outliers to test-response measurements, and only a sub-sample from the population is available, the distance metric becomes unreliable. To circumvent this problem, multiple Mahalanobis distances are calculated from selected sets of test-response measurements. They are then suitably formulated to derive a metric that has a reduced variance and robust to shifts or deviations in measurements. In this paper, such a formulation is proposed to qualitatively screen product outliers and quantitatively measure the reliability of the non-defective ones. The application of method is exemplified over a test set of an industrial automobile product.

    KW - METIS-284958

    KW - IR-79215

    KW - manufacturing test

    KW - EWI-21152

    KW - Metrics

    KW - CAES-TDT: Testable Design and Test

    KW - Outliers

    U2 - 10.1109/ETS.2011.31

    DO - 10.1109/ETS.2011.31

    M3 - Conference contribution

    SN - 978-1-4577-0483-3

    SP - 159

    EP - 164

    BT - Proceedings 16th IEEE European Test Symposium, ETS 2011

    PB - IEEE Computer Society

    CY - USA

    ER -

    Krishnan S, Krishnan S, Kerkhoff HG. A Robust Metric for Screening Outliers from Analogue Product Manufacturing Tests Responses. In Proceedings 16th IEEE European Test Symposium, ETS 2011. USA: IEEE Computer Society. 2011. p. 159-164 https://doi.org/10.1109/ETS.2011.31