Feature selection for portfolio optimization

Thomas Trier Bjerring, Omri Ross, Alex Weissensteiner

    Research output: Contribution to journalJournal articleResearchpeer-review

    645 Downloads (Pure)

    Abstract

    Most portfolio selection rules based on the sample mean and covariance matrix perform poorly out-of-sample. Moreover, there is a growing body of evidence that such optimization rules are not able to beat simple rules of thumb, such as 1/N. Parameter uncertainty has been identified as one major reason for these findings. A strand of literature addresses this problem by improving the parameter estimation and/or by relying on more robust portfolio selection methods. Independent of the chosen portfolio selection rule, we propose using feature selection first in order to reduce the asset menu. While most of the diversification benefits are preserved, the parameter estimation problem is alleviated. We conduct out-of-sample back-tests to show that in most cases different well-established portfolio selection rules applied on the reduced asset universe are able to improve alpha relative to different prominent factor models.
    Original languageEnglish
    JournalAnnals of Operations Research
    Volume256
    Issue number1
    Pages (from-to)21-40
    ISSN0254-5330
    DOIs
    Publication statusPublished - 2017

    Keywords

    • Agglomerative hierarchical clustering
    • Feature selection
    • Parameter uncertainty
    • Portfolio optimization

    Cite this

    Bjerring, Thomas Trier ; Ross, Omri ; Weissensteiner, Alex. / Feature selection for portfolio optimization. In: Annals of Operations Research. 2017 ; Vol. 256, No. 1. pp. 21-40.
    @article{666d40a899c94abfa42cafc697f451b4,
    title = "Feature selection for portfolio optimization",
    abstract = "Most portfolio selection rules based on the sample mean and covariance matrix perform poorly out-of-sample. Moreover, there is a growing body of evidence that such optimization rules are not able to beat simple rules of thumb, such as 1/N. Parameter uncertainty has been identified as one major reason for these findings. A strand of literature addresses this problem by improving the parameter estimation and/or by relying on more robust portfolio selection methods. Independent of the chosen portfolio selection rule, we propose using feature selection first in order to reduce the asset menu. While most of the diversification benefits are preserved, the parameter estimation problem is alleviated. We conduct out-of-sample back-tests to show that in most cases different well-established portfolio selection rules applied on the reduced asset universe are able to improve alpha relative to different prominent factor models.",
    keywords = "Agglomerative hierarchical clustering, Feature selection, Parameter uncertainty, Portfolio optimization",
    author = "Bjerring, {Thomas Trier} and Omri Ross and Alex Weissensteiner",
    year = "2017",
    doi = "10.1007/s10479-016-2155-y",
    language = "English",
    volume = "256",
    pages = "21--40",
    journal = "Annals of Operations Research",
    issn = "0254-5330",
    publisher = "Springer New York",
    number = "1",

    }

    Bjerring, TT, Ross, O & Weissensteiner, A 2017, 'Feature selection for portfolio optimization', Annals of Operations Research, vol. 256, no. 1, pp. 21-40. https://doi.org/10.1007/s10479-016-2155-y

    Feature selection for portfolio optimization. / Bjerring, Thomas Trier; Ross, Omri; Weissensteiner, Alex.

    In: Annals of Operations Research, Vol. 256, No. 1, 2017, p. 21-40.

    Research output: Contribution to journalJournal articleResearchpeer-review

    TY - JOUR

    T1 - Feature selection for portfolio optimization

    AU - Bjerring, Thomas Trier

    AU - Ross, Omri

    AU - Weissensteiner, Alex

    PY - 2017

    Y1 - 2017

    N2 - Most portfolio selection rules based on the sample mean and covariance matrix perform poorly out-of-sample. Moreover, there is a growing body of evidence that such optimization rules are not able to beat simple rules of thumb, such as 1/N. Parameter uncertainty has been identified as one major reason for these findings. A strand of literature addresses this problem by improving the parameter estimation and/or by relying on more robust portfolio selection methods. Independent of the chosen portfolio selection rule, we propose using feature selection first in order to reduce the asset menu. While most of the diversification benefits are preserved, the parameter estimation problem is alleviated. We conduct out-of-sample back-tests to show that in most cases different well-established portfolio selection rules applied on the reduced asset universe are able to improve alpha relative to different prominent factor models.

    AB - Most portfolio selection rules based on the sample mean and covariance matrix perform poorly out-of-sample. Moreover, there is a growing body of evidence that such optimization rules are not able to beat simple rules of thumb, such as 1/N. Parameter uncertainty has been identified as one major reason for these findings. A strand of literature addresses this problem by improving the parameter estimation and/or by relying on more robust portfolio selection methods. Independent of the chosen portfolio selection rule, we propose using feature selection first in order to reduce the asset menu. While most of the diversification benefits are preserved, the parameter estimation problem is alleviated. We conduct out-of-sample back-tests to show that in most cases different well-established portfolio selection rules applied on the reduced asset universe are able to improve alpha relative to different prominent factor models.

    KW - Agglomerative hierarchical clustering

    KW - Feature selection

    KW - Parameter uncertainty

    KW - Portfolio optimization

    U2 - 10.1007/s10479-016-2155-y

    DO - 10.1007/s10479-016-2155-y

    M3 - Journal article

    VL - 256

    SP - 21

    EP - 40

    JO - Annals of Operations Research

    JF - Annals of Operations Research

    SN - 0254-5330

    IS - 1

    ER -