Parametric spectral discrimination

Andrew J. Grant, Barry G. Quinn

    Research output: Contribution to journalArticleResearchpeer-review

    Abstract

    This article is concerned with determining whether two independent time series have been generated by underlying stochastic processes with the same spectral shape. There are many methods that do so using the periodogram. Alternative approaches test for the equality of a finite number of autocovariances or autocorrelations. Non-parametric methods usually have low power when compared with parametric methods. The parametric approach we introduce fits autoregressions to the two time series and tests whether the model parameters are equal using a likelihood ratio test. The test performs well when the time series are from autoregressions. However, problems arise when this is not the case. A modification to the test is proposed, which fits fixed order autoregressions. Simulations show that the modified test performs well even when the two time series are not from autoregressive processes. The parametric approach is shown to outperform non-parametric alternatives in a power study.

    LanguageEnglish
    Pages838-864
    Number of pages27
    JournalJournal of Time Series Analysis
    Volume38
    Issue number6
    Early online date24 Apr 2017
    DOIs
    Publication statusPublished - Nov 2017

    Fingerprint

    Discrimination
    Time series
    Autoregression
    Autocovariance
    Random processes
    Periodogram
    Autocorrelation
    Alternatives
    Nonparametric Methods
    Autoregressive Process
    Likelihood Ratio Test
    Stochastic Processes
    Equality
    Simulation
    Model

    Keywords

    • autoregression
    • discriminant analysis
    • spectral comparison
    • spectral density

    Cite this

    Grant, Andrew J. ; Quinn, Barry G. / Parametric spectral discrimination. In: Journal of Time Series Analysis. 2017 ; Vol. 38, No. 6. pp. 838-864.
    @article{890bc685aeca4e729689d45d837c3863,
    title = "Parametric spectral discrimination",
    abstract = "This article is concerned with determining whether two independent time series have been generated by underlying stochastic processes with the same spectral shape. There are many methods that do so using the periodogram. Alternative approaches test for the equality of a finite number of autocovariances or autocorrelations. Non-parametric methods usually have low power when compared with parametric methods. The parametric approach we introduce fits autoregressions to the two time series and tests whether the model parameters are equal using a likelihood ratio test. The test performs well when the time series are from autoregressions. However, problems arise when this is not the case. A modification to the test is proposed, which fits fixed order autoregressions. Simulations show that the modified test performs well even when the two time series are not from autoregressive processes. The parametric approach is shown to outperform non-parametric alternatives in a power study.",
    keywords = "autoregression, discriminant analysis, spectral comparison, spectral density",
    author = "Grant, {Andrew J.} and Quinn, {Barry G.}",
    year = "2017",
    month = "11",
    doi = "10.1111/jtsa.12238",
    language = "English",
    volume = "38",
    pages = "838--864",
    journal = "Journal of Time Series Analysis",
    issn = "0143-9782",
    publisher = "Wiley-Blackwell, Wiley",
    number = "6",

    }

    Parametric spectral discrimination. / Grant, Andrew J.; Quinn, Barry G.

    In: Journal of Time Series Analysis, Vol. 38, No. 6, 11.2017, p. 838-864.

    Research output: Contribution to journalArticleResearchpeer-review

    TY - JOUR

    T1 - Parametric spectral discrimination

    AU - Grant, Andrew J.

    AU - Quinn, Barry G.

    PY - 2017/11

    Y1 - 2017/11

    N2 - This article is concerned with determining whether two independent time series have been generated by underlying stochastic processes with the same spectral shape. There are many methods that do so using the periodogram. Alternative approaches test for the equality of a finite number of autocovariances or autocorrelations. Non-parametric methods usually have low power when compared with parametric methods. The parametric approach we introduce fits autoregressions to the two time series and tests whether the model parameters are equal using a likelihood ratio test. The test performs well when the time series are from autoregressions. However, problems arise when this is not the case. A modification to the test is proposed, which fits fixed order autoregressions. Simulations show that the modified test performs well even when the two time series are not from autoregressive processes. The parametric approach is shown to outperform non-parametric alternatives in a power study.

    AB - This article is concerned with determining whether two independent time series have been generated by underlying stochastic processes with the same spectral shape. There are many methods that do so using the periodogram. Alternative approaches test for the equality of a finite number of autocovariances or autocorrelations. Non-parametric methods usually have low power when compared with parametric methods. The parametric approach we introduce fits autoregressions to the two time series and tests whether the model parameters are equal using a likelihood ratio test. The test performs well when the time series are from autoregressions. However, problems arise when this is not the case. A modification to the test is proposed, which fits fixed order autoregressions. Simulations show that the modified test performs well even when the two time series are not from autoregressive processes. The parametric approach is shown to outperform non-parametric alternatives in a power study.

    KW - autoregression

    KW - discriminant analysis

    KW - spectral comparison

    KW - spectral density

    UR - http://www.scopus.com/inward/record.url?scp=85018877721&partnerID=8YFLogxK

    U2 - 10.1111/jtsa.12238

    DO - 10.1111/jtsa.12238

    M3 - Article

    VL - 38

    SP - 838

    EP - 864

    JO - Journal of Time Series Analysis

    T2 - Journal of Time Series Analysis

    JF - Journal of Time Series Analysis

    SN - 0143-9782

    IS - 6

    ER -