Robust Multivariate L1 Principal Component Analysis and Dimensionality Reduction

Junbin Gao, Paul Kwan, Yi Guo

    Research output: Contribution to journalArticlepeer-review

    12 Citations (Scopus)
    104 Downloads (Pure)

    Abstract

    Further to our recent work on the robust L1 PCA we introduce a new ver-sion of robust PCA model based on the so-called multivariate Laplace distribution(called L1 distribution) proposed in (Eltoft et al., 2006). Due to the heavy tail and high component dependency characteristics of the multivariate L1 distribution, the proposed model is expected to be more robust against data outliers and ¯ttingcomponent dependency. Additionally, we demonstrate how a variational approx-imation scheme enables e®ective inference of key parameters in the probabilistic multivariate L1-PCA model. By doing so, a tractable Bayesian inference can be achieved based on the variational EM-type algorithm.
    Original languageEnglish
    Pages (from-to)1242-1249
    Number of pages8
    JournalNeurocomputing
    Volume72
    Issue number4-6
    DOIs
    Publication statusPublished - 2009

    Fingerprint

    Dive into the research topics of 'Robust Multivariate L1 Principal Component Analysis and Dimensionality Reduction'. Together they form a unique fingerprint.

    Cite this