Low-rank dimensionality reduction for multi-modality neurodegenerative disease identification

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

In this paper, we propose a novel dimensionality reduction method of taking the advantages of the variability, sparsity, and low-rankness of neuroimaging data for Alzheimer’s Disease (AD) classification. We first take the variability of neuroimaging data into account by partitioning them into sub-classes by means of clustering, which thus captures the underlying multi-peak distributional characteristics in neuroimaging data. We then iteratively conduct Low-Rank Dimensionality Reduction (LRDR) and orthogonal rotation in a sparse linear regression framework, in order to find the low-dimensional structure of high-dimensional data. Experimental results on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset showed that our proposed model helped enhance the performances of AD classification, outperforming the state-of-the-art methods.

Original languageEnglish
JournalWorld Wide Web
DOIs
Publication statusAccepted/In press - 2018 Jan 1

Fingerprint

Neurodegenerative diseases
Neuroimaging
Linear regression

Keywords

  • Alzheimer’s Disease (AD)
  • Feature selection
  • Subspace learning

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

@article{3dbf97ae5344402fa6a3fc2759383c61,
title = "Low-rank dimensionality reduction for multi-modality neurodegenerative disease identification",
abstract = "In this paper, we propose a novel dimensionality reduction method of taking the advantages of the variability, sparsity, and low-rankness of neuroimaging data for Alzheimer’s Disease (AD) classification. We first take the variability of neuroimaging data into account by partitioning them into sub-classes by means of clustering, which thus captures the underlying multi-peak distributional characteristics in neuroimaging data. We then iteratively conduct Low-Rank Dimensionality Reduction (LRDR) and orthogonal rotation in a sparse linear regression framework, in order to find the low-dimensional structure of high-dimensional data. Experimental results on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset showed that our proposed model helped enhance the performances of AD classification, outperforming the state-of-the-art methods.",
keywords = "Alzheimer’s Disease (AD), Feature selection, Subspace learning",
author = "Xiaofeng Zhu and Heung-Il Suk and Dinggang Shen",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/s11280-018-0645-3",
language = "English",
journal = "World Wide Web",
issn = "1386-145X",
publisher = "Springer New York",

}

TY - JOUR

T1 - Low-rank dimensionality reduction for multi-modality neurodegenerative disease identification

AU - Zhu, Xiaofeng

AU - Suk, Heung-Il

AU - Shen, Dinggang

PY - 2018/1/1

Y1 - 2018/1/1

N2 - In this paper, we propose a novel dimensionality reduction method of taking the advantages of the variability, sparsity, and low-rankness of neuroimaging data for Alzheimer’s Disease (AD) classification. We first take the variability of neuroimaging data into account by partitioning them into sub-classes by means of clustering, which thus captures the underlying multi-peak distributional characteristics in neuroimaging data. We then iteratively conduct Low-Rank Dimensionality Reduction (LRDR) and orthogonal rotation in a sparse linear regression framework, in order to find the low-dimensional structure of high-dimensional data. Experimental results on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset showed that our proposed model helped enhance the performances of AD classification, outperforming the state-of-the-art methods.

AB - In this paper, we propose a novel dimensionality reduction method of taking the advantages of the variability, sparsity, and low-rankness of neuroimaging data for Alzheimer’s Disease (AD) classification. We first take the variability of neuroimaging data into account by partitioning them into sub-classes by means of clustering, which thus captures the underlying multi-peak distributional characteristics in neuroimaging data. We then iteratively conduct Low-Rank Dimensionality Reduction (LRDR) and orthogonal rotation in a sparse linear regression framework, in order to find the low-dimensional structure of high-dimensional data. Experimental results on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset showed that our proposed model helped enhance the performances of AD classification, outperforming the state-of-the-art methods.

KW - Alzheimer’s Disease (AD)

KW - Feature selection

KW - Subspace learning

UR - http://www.scopus.com/inward/record.url?scp=85056463714&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85056463714&partnerID=8YFLogxK

U2 - 10.1007/s11280-018-0645-3

DO - 10.1007/s11280-018-0645-3

M3 - Article

AN - SCOPUS:85056463714

JO - World Wide Web

JF - World Wide Web

SN - 1386-145X

ER -