On the evaluation of a nonlinear Principal Component Analysis

Ryo Saegusa, Shuji Hashimoto

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

    2 Citations (Scopus)

    Abstract

    Principal Component Analysis (PCA) is a useful method in multivariate analysis to reduce the dimensionality of data. We have already proposed a non-linearly extended model of PCA by employing neural networks and have shown its effectiveness with some artificial data. In this paper, we report results of a nonlinear principal component analysis on real-world data utilizing the proposed method. Moreover, we compare the distribution of reconstructed data with the distribution of the original data to discuss the advantage of nonlinear PCA.

    Original languageEnglish
    Title of host publicationProceedings of the IASTED International Conference on Neural Networks and Computational Intelligence
    EditorsM.H. Hamza
    Pages66-72
    Number of pages7
    Publication statusPublished - 2004
    EventProceedings of the IASTED International Conference on Neural Networks and Computational Intelligence - Grindelwald
    Duration: 2004 Feb 232004 Feb 25

    Other

    OtherProceedings of the IASTED International Conference on Neural Networks and Computational Intelligence
    CityGrindelwald
    Period04/2/2304/2/25

    Keywords

    • Dimensionality reduction
    • MLP
    • Nonlinear PCA

    ASJC Scopus subject areas

    • Engineering(all)

    Fingerprint Dive into the research topics of 'On the evaluation of a nonlinear Principal Component Analysis'. Together they form a unique fingerprint.

  • Cite this

    Saegusa, R., & Hashimoto, S. (2004). On the evaluation of a nonlinear Principal Component Analysis. In M. H. Hamza (Ed.), Proceedings of the IASTED International Conference on Neural Networks and Computational Intelligence (pp. 66-72). [413-084]