Nonlinear principal component analysis to preserve the order of principal components

Ryo Saegusa, Hitoshi Sakano, Shuji Hashimoto

    Research output: Contribution to journalArticle

    44 Citations (Scopus)

    Abstract

    Principal component analysis (PCA) is an effective method of linear dimensional reduction. Because of its simplicity in theory and implementation, it is often used for analyses in various disciplines. However, because of its linearity, PCA is not always suitable, and has redundancy in expressing data. To overcome this problem, some nonlinear PCA methods have been proposed. However, most of these methods have drawbacks, such that the number of principal components must be predetermined, and also the order of the generated principal components is not explicitly given. In this paper, we propose a nonlinear PCA algorithm that nonlinearly transforms data into principal components, and at the same time, preserving the order of the principal components, and we also propose a hierarchical neural network model to perform the algorithm. Moreover, our method does not need to know the number of principal components in advance. The effectiveness of the proposed model will be shown through experiments.

    Original languageEnglish
    Pages (from-to)57-70
    Number of pages14
    JournalNeurocomputing
    Volume61
    Issue number1-4
    DOIs
    Publication statusPublished - 2004 Oct

    Fingerprint

    Principal Component Analysis
    Principal component analysis
    Neural Networks (Computer)
    Redundancy
    Neural networks
    Experiments

    Keywords

    • Hierarchical structure
    • Nonlinear principal component analysis
    • Sand-glass type multi-layered perceptron

    ASJC Scopus subject areas

    • Artificial Intelligence
    • Cellular and Molecular Neuroscience

    Cite this

    Nonlinear principal component analysis to preserve the order of principal components. / Saegusa, Ryo; Sakano, Hitoshi; Hashimoto, Shuji.

    In: Neurocomputing, Vol. 61, No. 1-4, 10.2004, p. 57-70.

    Research output: Contribution to journalArticle

    Saegusa, Ryo ; Sakano, Hitoshi ; Hashimoto, Shuji. / Nonlinear principal component analysis to preserve the order of principal components. In: Neurocomputing. 2004 ; Vol. 61, No. 1-4. pp. 57-70.
    @article{4306ebb68a8f416280b36e3741722fc3,
    title = "Nonlinear principal component analysis to preserve the order of principal components",
    abstract = "Principal component analysis (PCA) is an effective method of linear dimensional reduction. Because of its simplicity in theory and implementation, it is often used for analyses in various disciplines. However, because of its linearity, PCA is not always suitable, and has redundancy in expressing data. To overcome this problem, some nonlinear PCA methods have been proposed. However, most of these methods have drawbacks, such that the number of principal components must be predetermined, and also the order of the generated principal components is not explicitly given. In this paper, we propose a nonlinear PCA algorithm that nonlinearly transforms data into principal components, and at the same time, preserving the order of the principal components, and we also propose a hierarchical neural network model to perform the algorithm. Moreover, our method does not need to know the number of principal components in advance. The effectiveness of the proposed model will be shown through experiments.",
    keywords = "Hierarchical structure, Nonlinear principal component analysis, Sand-glass type multi-layered perceptron",
    author = "Ryo Saegusa and Hitoshi Sakano and Shuji Hashimoto",
    year = "2004",
    month = "10",
    doi = "10.1016/j.neucom.2004.03.004",
    language = "English",
    volume = "61",
    pages = "57--70",
    journal = "Neurocomputing",
    issn = "0925-2312",
    publisher = "Elsevier",
    number = "1-4",

    }

    TY - JOUR

    T1 - Nonlinear principal component analysis to preserve the order of principal components

    AU - Saegusa, Ryo

    AU - Sakano, Hitoshi

    AU - Hashimoto, Shuji

    PY - 2004/10

    Y1 - 2004/10

    N2 - Principal component analysis (PCA) is an effective method of linear dimensional reduction. Because of its simplicity in theory and implementation, it is often used for analyses in various disciplines. However, because of its linearity, PCA is not always suitable, and has redundancy in expressing data. To overcome this problem, some nonlinear PCA methods have been proposed. However, most of these methods have drawbacks, such that the number of principal components must be predetermined, and also the order of the generated principal components is not explicitly given. In this paper, we propose a nonlinear PCA algorithm that nonlinearly transforms data into principal components, and at the same time, preserving the order of the principal components, and we also propose a hierarchical neural network model to perform the algorithm. Moreover, our method does not need to know the number of principal components in advance. The effectiveness of the proposed model will be shown through experiments.

    AB - Principal component analysis (PCA) is an effective method of linear dimensional reduction. Because of its simplicity in theory and implementation, it is often used for analyses in various disciplines. However, because of its linearity, PCA is not always suitable, and has redundancy in expressing data. To overcome this problem, some nonlinear PCA methods have been proposed. However, most of these methods have drawbacks, such that the number of principal components must be predetermined, and also the order of the generated principal components is not explicitly given. In this paper, we propose a nonlinear PCA algorithm that nonlinearly transforms data into principal components, and at the same time, preserving the order of the principal components, and we also propose a hierarchical neural network model to perform the algorithm. Moreover, our method does not need to know the number of principal components in advance. The effectiveness of the proposed model will be shown through experiments.

    KW - Hierarchical structure

    KW - Nonlinear principal component analysis

    KW - Sand-glass type multi-layered perceptron

    UR - http://www.scopus.com/inward/record.url?scp=10244247743&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=10244247743&partnerID=8YFLogxK

    U2 - 10.1016/j.neucom.2004.03.004

    DO - 10.1016/j.neucom.2004.03.004

    M3 - Article

    AN - SCOPUS:10244247743

    VL - 61

    SP - 57

    EP - 70

    JO - Neurocomputing

    JF - Neurocomputing

    SN - 0925-2312

    IS - 1-4

    ER -