A proposal of l<inf>1</inf> regularized distance metric learning for high dimensional sparse vector space

Kenta Mikawa, Manabu Kobayashi, Masayuki Goto, Shigeichi Hirasawa

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    1 Citation (Scopus)

    Abstract

    In this paper, we focus on pattern recognition based on the vector space model with the high dimensional and sparse data. One of the pattern recognition methods is metric learning which learns a metric matrix by using the iterative optimization procedure. However most of the metric learning methods tend to cause overfitting and increasing computational time for high dimensional and sparse settings. To avoid these problems, we propose the method of l<inf>1</inf> regularized metric learning by using the algorithm of alternating direction method of multiplier (ADMM) in the supervised setting. The effectiveness of our proposed method is clarified by classification experiments by using the Japanese newspaper article and UCI machine learning repository. And we show proposed method is the special case of the statistical sparse covariance selection.

    Original languageEnglish
    Title of host publicationConference Proceedings - IEEE International Conference on Systems, Man and Cybernetics
    PublisherInstitute of Electrical and Electronics Engineers Inc.
    Pages1985-1990
    Number of pages6
    Volume2014-January
    EditionJanuary
    DOIs
    Publication statusPublished - 2014
    Event2014 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2014 - San Diego, United States
    Duration: 2014 Oct 52014 Oct 8

    Other

    Other2014 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2014
    CountryUnited States
    CitySan Diego
    Period14/10/514/10/8

    Fingerprint

    Vector spaces
    Pattern recognition
    Learning systems
    Experiments

    ASJC Scopus subject areas

    • Electrical and Electronic Engineering
    • Control and Systems Engineering
    • Human-Computer Interaction

    Cite this

    Mikawa, K., Kobayashi, M., Goto, M., & Hirasawa, S. (2014). A proposal of l<inf>1</inf> regularized distance metric learning for high dimensional sparse vector space. In Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics (January ed., Vol. 2014-January, pp. 1985-1990). [6974212] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/smc.2014.6974212

    A proposal of l<inf>1</inf> regularized distance metric learning for high dimensional sparse vector space. / Mikawa, Kenta; Kobayashi, Manabu; Goto, Masayuki; Hirasawa, Shigeichi.

    Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics. Vol. 2014-January January. ed. Institute of Electrical and Electronics Engineers Inc., 2014. p. 1985-1990 6974212.

    Research output: Chapter in Book/Report/Conference proceedingConference contribution

    Mikawa, K, Kobayashi, M, Goto, M & Hirasawa, S 2014, A proposal of l<inf>1</inf> regularized distance metric learning for high dimensional sparse vector space. in Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics. January edn, vol. 2014-January, 6974212, Institute of Electrical and Electronics Engineers Inc., pp. 1985-1990, 2014 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2014, San Diego, United States, 14/10/5. https://doi.org/10.1109/smc.2014.6974212
    Mikawa K, Kobayashi M, Goto M, Hirasawa S. A proposal of l<inf>1</inf> regularized distance metric learning for high dimensional sparse vector space. In Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics. January ed. Vol. 2014-January. Institute of Electrical and Electronics Engineers Inc. 2014. p. 1985-1990. 6974212 https://doi.org/10.1109/smc.2014.6974212
    Mikawa, Kenta ; Kobayashi, Manabu ; Goto, Masayuki ; Hirasawa, Shigeichi. / A proposal of l<inf>1</inf> regularized distance metric learning for high dimensional sparse vector space. Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics. Vol. 2014-January January. ed. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 1985-1990
    @inproceedings{081b88b139c84c88b79ddff289e1bb1d,
    title = "A proposal of l1 regularized distance metric learning for high dimensional sparse vector space",
    abstract = "In this paper, we focus on pattern recognition based on the vector space model with the high dimensional and sparse data. One of the pattern recognition methods is metric learning which learns a metric matrix by using the iterative optimization procedure. However most of the metric learning methods tend to cause overfitting and increasing computational time for high dimensional and sparse settings. To avoid these problems, we propose the method of l1 regularized metric learning by using the algorithm of alternating direction method of multiplier (ADMM) in the supervised setting. The effectiveness of our proposed method is clarified by classification experiments by using the Japanese newspaper article and UCI machine learning repository. And we show proposed method is the special case of the statistical sparse covariance selection.",
    author = "Kenta Mikawa and Manabu Kobayashi and Masayuki Goto and Shigeichi Hirasawa",
    year = "2014",
    doi = "10.1109/smc.2014.6974212",
    language = "English",
    volume = "2014-January",
    pages = "1985--1990",
    booktitle = "Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics",
    publisher = "Institute of Electrical and Electronics Engineers Inc.",
    edition = "January",

    }

    TY - GEN

    T1 - A proposal of l1 regularized distance metric learning for high dimensional sparse vector space

    AU - Mikawa, Kenta

    AU - Kobayashi, Manabu

    AU - Goto, Masayuki

    AU - Hirasawa, Shigeichi

    PY - 2014

    Y1 - 2014

    N2 - In this paper, we focus on pattern recognition based on the vector space model with the high dimensional and sparse data. One of the pattern recognition methods is metric learning which learns a metric matrix by using the iterative optimization procedure. However most of the metric learning methods tend to cause overfitting and increasing computational time for high dimensional and sparse settings. To avoid these problems, we propose the method of l1 regularized metric learning by using the algorithm of alternating direction method of multiplier (ADMM) in the supervised setting. The effectiveness of our proposed method is clarified by classification experiments by using the Japanese newspaper article and UCI machine learning repository. And we show proposed method is the special case of the statistical sparse covariance selection.

    AB - In this paper, we focus on pattern recognition based on the vector space model with the high dimensional and sparse data. One of the pattern recognition methods is metric learning which learns a metric matrix by using the iterative optimization procedure. However most of the metric learning methods tend to cause overfitting and increasing computational time for high dimensional and sparse settings. To avoid these problems, we propose the method of l1 regularized metric learning by using the algorithm of alternating direction method of multiplier (ADMM) in the supervised setting. The effectiveness of our proposed method is clarified by classification experiments by using the Japanese newspaper article and UCI machine learning repository. And we show proposed method is the special case of the statistical sparse covariance selection.

    UR - http://www.scopus.com/inward/record.url?scp=84938059660&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=84938059660&partnerID=8YFLogxK

    U2 - 10.1109/smc.2014.6974212

    DO - 10.1109/smc.2014.6974212

    M3 - Conference contribution

    AN - SCOPUS:84938059660

    VL - 2014-January

    SP - 1985

    EP - 1990

    BT - Conference Proceedings - IEEE International Conference on Systems, Man and Cybernetics

    PB - Institute of Electrical and Electronics Engineers Inc.

    ER -