Block tensor train decomposition for missing data estimation

Namgil Lee, Jong Min Kim

    Research output: Contribution to journalArticle

    1 Scopus citations

    Abstract

    We propose a method for imputation of missing values in large scale matrix data based on a low-rank tensor approximation technique called the block tensor train (BTT) decomposition. Given sparsely observed data points, the proposed method iteratively computes the singular value decomposition (SVD) of the underlying data matrix with missing values. The SVD of the matrices is performed based on a low-rank BTT decomposition, by which storage and time complexities can be reduced dramatically for large-scale data matrices admitting a low-rank tensor structure. An iterative soft-thresholding algorithm is implemented for missing data estimation based on an alternating least squares method for BTT decomposition. Experimental results on simulated data and real benchmark data demonstrate that the proposed method can estimate a large amount of missing values accurately compared to a matrix-based standard method. The R source code of the BTT-based imputation method is available at https://github.com/namgillee/BTTSoftImpute.

    Original languageEnglish (US)
    Pages (from-to)1283-1305
    Number of pages23
    JournalStatistical Papers
    Volume59
    Issue number4
    DOIs
    StatePublished - Dec 1 2018

    Keywords

    • Imputation
    • Multidimensional array
    • Singular value decomposition
    • Tensor network

    Fingerprint Dive into the research topics of 'Block tensor train decomposition for missing data estimation'. Together they form a unique fingerprint.

  • Cite this