Robust simultaneous low rank approximation of tensors

Kohei Inoue, Hara Kenji, Kiichi Urahama

Research output: Contribution to journalConference article

2 Citations (Scopus)

Abstract

We propose simultaneous low rank approximation of tensors (SLRAT) for the dimensionality reduction of tensors and modify it to the robust one, i.e., the robust SLRAT. For both the SLRAT and the robust SLRAT, we propose iterative algorithms for solving them. It is experimentally shown that the robust SLRAT achieves lower reconstruction error than the SLRAT when a dataset contains noise data. We also propose a method for classifying sets of tensors and call it the subspace matching, where both training data and testing data are represented by their subspaces, and each testing datum is classified on the basis of the similarity between subspaces. It is experimentally verified that the robust SLRAT achieves higher recognition rate than the SLRAT when the testing data contain noise data.

Original languageEnglish
Pages (from-to)574-584
Number of pages11
JournalLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume5414 LNCS
DOIs
Publication statusPublished - Feb 19 2009
Event3rd Pacific Rim Symposium on Image and Video Technology, PSIVT 2009 - Tokyo, Japan
Duration: Jan 13 2009Jan 16 2009

Fingerprint

Low-rank Approximation
Simultaneous Approximation
Tensors
Tensor
Subspace
Testing
Dimensionality Reduction
Iterative Algorithm

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

@article{d854b0654e51441891563ea2fa3d5833,
title = "Robust simultaneous low rank approximation of tensors",
abstract = "We propose simultaneous low rank approximation of tensors (SLRAT) for the dimensionality reduction of tensors and modify it to the robust one, i.e., the robust SLRAT. For both the SLRAT and the robust SLRAT, we propose iterative algorithms for solving them. It is experimentally shown that the robust SLRAT achieves lower reconstruction error than the SLRAT when a dataset contains noise data. We also propose a method for classifying sets of tensors and call it the subspace matching, where both training data and testing data are represented by their subspaces, and each testing datum is classified on the basis of the similarity between subspaces. It is experimentally verified that the robust SLRAT achieves higher recognition rate than the SLRAT when the testing data contain noise data.",
author = "Kohei Inoue and Hara Kenji and Kiichi Urahama",
year = "2009",
month = "2",
day = "19",
doi = "10.1007/978-3-540-92957-4_50",
language = "English",
volume = "5414 LNCS",
pages = "574--584",
journal = "Lecture Notes in Computer Science",
issn = "0302-9743",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - Robust simultaneous low rank approximation of tensors

AU - Inoue, Kohei

AU - Kenji, Hara

AU - Urahama, Kiichi

PY - 2009/2/19

Y1 - 2009/2/19

N2 - We propose simultaneous low rank approximation of tensors (SLRAT) for the dimensionality reduction of tensors and modify it to the robust one, i.e., the robust SLRAT. For both the SLRAT and the robust SLRAT, we propose iterative algorithms for solving them. It is experimentally shown that the robust SLRAT achieves lower reconstruction error than the SLRAT when a dataset contains noise data. We also propose a method for classifying sets of tensors and call it the subspace matching, where both training data and testing data are represented by their subspaces, and each testing datum is classified on the basis of the similarity between subspaces. It is experimentally verified that the robust SLRAT achieves higher recognition rate than the SLRAT when the testing data contain noise data.

AB - We propose simultaneous low rank approximation of tensors (SLRAT) for the dimensionality reduction of tensors and modify it to the robust one, i.e., the robust SLRAT. For both the SLRAT and the robust SLRAT, we propose iterative algorithms for solving them. It is experimentally shown that the robust SLRAT achieves lower reconstruction error than the SLRAT when a dataset contains noise data. We also propose a method for classifying sets of tensors and call it the subspace matching, where both training data and testing data are represented by their subspaces, and each testing datum is classified on the basis of the similarity between subspaces. It is experimentally verified that the robust SLRAT achieves higher recognition rate than the SLRAT when the testing data contain noise data.

UR - http://www.scopus.com/inward/record.url?scp=60149109506&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=60149109506&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-92957-4_50

DO - 10.1007/978-3-540-92957-4_50

M3 - Conference article

AN - SCOPUS:60149109506

VL - 5414 LNCS

SP - 574

EP - 584

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

SN - 0302-9743

ER -