Patchworking multiple pairwise distances for learning with distance matrices

Ken Takano, Hideitsu Hino, Yuki Yoshikawa, Noboru Murata

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

    1 Citation (Scopus)

    Abstract

    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.

    Original languageEnglish
    Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    PublisherSpringer Verlag
    Pages287-294
    Number of pages8
    Volume9237
    ISBN (Print)9783319224817
    DOIs
    Publication statusPublished - 2015
    Event12th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2015 - Liberec, Czech Republic
    Duration: 2015 Aug 252015 Aug 28

    Publication series

    NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
    Volume9237
    ISSN (Print)03029743
    ISSN (Electronic)16113349

    Other

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

    Keywords

    • Classification
    • Decision trees
    • Graph kernel
    • Random forest
    • Spike train
    • Structured data

    ASJC Scopus subject areas

    • Computer Science(all)
    • Theoretical Computer Science

    Fingerprint Dive into the research topics of 'Patchworking multiple pairwise distances for learning with distance matrices'. Together they form a unique fingerprint.

  • Cite this

    Takano, K., Hino, H., Yoshikawa, Y., & Murata, N. (2015). Patchworking multiple pairwise distances for learning with distance matrices. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9237, pp. 287-294). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9237). Springer Verlag. https://doi.org/10.1007/978-3-319-22482-4_33