Low-rank and sparse matrix factorization for scientific paper recommendation in heterogeneous network

Tao Dai, Tianyu Gao, Li Zhu, Xiaoyan Cai, Shirui Pan

Research output: Contribution to journalArticleResearchpeer-review

Abstract

With the rapid growth of scientific publications, it is hard for researchers to acquire appropriate papers that meet their expectations. Recommendation system for scientific articles is an essential technology to overcome this problem. In this paper, we propose a novel low-rank and sparse matrix factorization-based paper recommendation (LSMFPRec) method for authors. The proposed method seamlessly combines low-rank and sparse matrix factorization method with fine-grained paper and author affinity matrixes that are extracted from heterogeneous scientific network. Thus, it can effectively alleviate the sparsity and cold start problems that exist in traditional matrix factorization based collaborative filtering methods. Moreover, LSMFPRec can significantly reduce the error propagated from intermediate outputs. In addition, the proposed method essentially captures the low-rank and sparse characteristics that exist in scientific rating activities; therefore, it can generate more reasonable predicted ratings for influential and uninfluential papers. The effectiveness of the proposed LSMFPRec is demonstrated by the recommendation evaluation conducted on the AAN and CiteULike data sets.

Original languageEnglish
Article number8434216
Pages (from-to)59015-59030
Number of pages16
JournalIEEE Access
Volume6
DOIs
Publication statusPublished - 13 Aug 2018
Externally publishedYes

Keywords

  • Heterogeneous network
  • Low rank and sparse matrix factorization
  • Paper recommendation

Cite this

Dai, Tao ; Gao, Tianyu ; Zhu, Li ; Cai, Xiaoyan ; Pan, Shirui. / Low-rank and sparse matrix factorization for scientific paper recommendation in heterogeneous network. In: IEEE Access. 2018 ; Vol. 6. pp. 59015-59030.
@article{53d931c24626442fa3b827c74bffe24b,
title = "Low-rank and sparse matrix factorization for scientific paper recommendation in heterogeneous network",
abstract = "With the rapid growth of scientific publications, it is hard for researchers to acquire appropriate papers that meet their expectations. Recommendation system for scientific articles is an essential technology to overcome this problem. In this paper, we propose a novel low-rank and sparse matrix factorization-based paper recommendation (LSMFPRec) method for authors. The proposed method seamlessly combines low-rank and sparse matrix factorization method with fine-grained paper and author affinity matrixes that are extracted from heterogeneous scientific network. Thus, it can effectively alleviate the sparsity and cold start problems that exist in traditional matrix factorization based collaborative filtering methods. Moreover, LSMFPRec can significantly reduce the error propagated from intermediate outputs. In addition, the proposed method essentially captures the low-rank and sparse characteristics that exist in scientific rating activities; therefore, it can generate more reasonable predicted ratings for influential and uninfluential papers. The effectiveness of the proposed LSMFPRec is demonstrated by the recommendation evaluation conducted on the AAN and CiteULike data sets.",
keywords = "Heterogeneous network, Low rank and sparse matrix factorization, Paper recommendation",
author = "Tao Dai and Tianyu Gao and Li Zhu and Xiaoyan Cai and Shirui Pan",
year = "2018",
month = "8",
day = "13",
doi = "10.1109/ACCESS.2018.2865115",
language = "English",
volume = "6",
pages = "59015--59030",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",

}

Low-rank and sparse matrix factorization for scientific paper recommendation in heterogeneous network. / Dai, Tao; Gao, Tianyu; Zhu, Li; Cai, Xiaoyan; Pan, Shirui.

In: IEEE Access, Vol. 6, 8434216, 13.08.2018, p. 59015-59030.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Low-rank and sparse matrix factorization for scientific paper recommendation in heterogeneous network

AU - Dai, Tao

AU - Gao, Tianyu

AU - Zhu, Li

AU - Cai, Xiaoyan

AU - Pan, Shirui

PY - 2018/8/13

Y1 - 2018/8/13

N2 - With the rapid growth of scientific publications, it is hard for researchers to acquire appropriate papers that meet their expectations. Recommendation system for scientific articles is an essential technology to overcome this problem. In this paper, we propose a novel low-rank and sparse matrix factorization-based paper recommendation (LSMFPRec) method for authors. The proposed method seamlessly combines low-rank and sparse matrix factorization method with fine-grained paper and author affinity matrixes that are extracted from heterogeneous scientific network. Thus, it can effectively alleviate the sparsity and cold start problems that exist in traditional matrix factorization based collaborative filtering methods. Moreover, LSMFPRec can significantly reduce the error propagated from intermediate outputs. In addition, the proposed method essentially captures the low-rank and sparse characteristics that exist in scientific rating activities; therefore, it can generate more reasonable predicted ratings for influential and uninfluential papers. The effectiveness of the proposed LSMFPRec is demonstrated by the recommendation evaluation conducted on the AAN and CiteULike data sets.

AB - With the rapid growth of scientific publications, it is hard for researchers to acquire appropriate papers that meet their expectations. Recommendation system for scientific articles is an essential technology to overcome this problem. In this paper, we propose a novel low-rank and sparse matrix factorization-based paper recommendation (LSMFPRec) method for authors. The proposed method seamlessly combines low-rank and sparse matrix factorization method with fine-grained paper and author affinity matrixes that are extracted from heterogeneous scientific network. Thus, it can effectively alleviate the sparsity and cold start problems that exist in traditional matrix factorization based collaborative filtering methods. Moreover, LSMFPRec can significantly reduce the error propagated from intermediate outputs. In addition, the proposed method essentially captures the low-rank and sparse characteristics that exist in scientific rating activities; therefore, it can generate more reasonable predicted ratings for influential and uninfluential papers. The effectiveness of the proposed LSMFPRec is demonstrated by the recommendation evaluation conducted on the AAN and CiteULike data sets.

KW - Heterogeneous network

KW - Low rank and sparse matrix factorization

KW - Paper recommendation

UR - http://www.scopus.com/inward/record.url?scp=85051652519&partnerID=8YFLogxK

U2 - 10.1109/ACCESS.2018.2865115

DO - 10.1109/ACCESS.2018.2865115

M3 - Article

VL - 6

SP - 59015

EP - 59030

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

M1 - 8434216

ER -