Patchworking multiple pairwise distances for learning with distance matrices

Ken Takano, Hideitsu Hino*, Yuki Yoshikawa, Noboru Murata

*この研究の対応する著者

    研究成果: Conference contribution

    1 被引用数 (Scopus)

    抄録

    A classification framework using only a set of distance matrices is proposed. The proposed algorithm can learn a classifier only from a set of distance matrices or similarity matrices, hence applicable to structured data, which do not have natural vector representation such as time series and graphs. Random forest is used to explore ideal feature representation based on the distance between points defined by a set of given distance matrices. The effectiveness of the proposed method is evaluated through experiments with point process data and graph structured data.

    本文言語English
    ホスト出版物のタイトルLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    出版社Springer Verlag
    ページ287-294
    ページ数8
    9237
    ISBN(印刷版)9783319224817
    DOI
    出版ステータスPublished - 2015
    イベント12th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2015 - Liberec, Czech Republic
    継続期間: 2015 8月 252015 8月 28

    出版物シリーズ

    名前Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    9237
    ISSN(印刷版)03029743
    ISSN(電子版)16113349

    Other

    Other12th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2015
    国/地域Czech Republic
    CityLiberec
    Period15/8/2515/8/28

    ASJC Scopus subject areas

    • コンピュータ サイエンス(全般)
    • 理論的コンピュータサイエンス

    フィンガープリント

    「Patchworking multiple pairwise distances for learning with distance matrices」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

    引用スタイル