Coupled Graph and Tensor Factorization for Recommender Systems and Community Detection

Vassilis N. Ioannidis, Ahmed S. Zamzam, Georgios B. Giannakis, Nicholas D. Sidiropoulos

Research output: Contribution to journalArticle

Abstract

Joint analysis of data from multiple information repositories facilitates uncovering the underlying structure in heterogeneous datasets. Single and coupled matrix-tensor factorization (CMTF) has been widely used in this context for imputation-based recommendation from ratings, social network, and other user-item data. When this side information is in the form of item-item correlation matrices or graphs, existing CMTF algorithms may fall short. Alleviating current limitations, we introduce a novel model coined coupled graph-tensor factorization (CGTF) that judiciously accounts for graph-related side information. The CGTF model has the potential to overcome practical challenges, such as missing slabs from the tensor and/or missing rows/columns from the correlation matrices. A novel alternating direction method of multipliers (ADMM) is also developed that recovers the nonnegative factors of CGTF. Our algorithm enjoys closed-form updates that result in reduced computational complexity and allow for convergence claims. A novel direction is further explored by employing the interpretable factors to detect graph communities having the tensor as side information. The resulting community detection approach is successful even when some links in the graphs are missing. Results with real data sets corroborate the merits of the proposed methods relative to state-of-the-art competing factorization techniques in providing recommendations and detecting communities.

Original languageEnglish (US)
JournalIEEE Transactions on Knowledge and Data Engineering
DOIs
StateAccepted/In press - Jan 1 2019

Fingerprint

Recommender systems
Factorization
Tensors
Computational complexity

Keywords

  • Tensor-matrix factorization
  • community detection
  • graph data
  • recommender systems
  • tensor-graph imputation

Cite this

Coupled Graph and Tensor Factorization for Recommender Systems and Community Detection. / Ioannidis, Vassilis N.; Zamzam, Ahmed S.; Giannakis, Georgios B.; Sidiropoulos, Nicholas D.

In: IEEE Transactions on Knowledge and Data Engineering, 01.01.2019.

Research output: Contribution to journalArticle

@article{b9138af3df0b4f719adac0d5004442fa,
title = "Coupled Graph and Tensor Factorization for Recommender Systems and Community Detection",
abstract = "Joint analysis of data from multiple information repositories facilitates uncovering the underlying structure in heterogeneous datasets. Single and coupled matrix-tensor factorization (CMTF) has been widely used in this context for imputation-based recommendation from ratings, social network, and other user-item data. When this side information is in the form of item-item correlation matrices or graphs, existing CMTF algorithms may fall short. Alleviating current limitations, we introduce a novel model coined coupled graph-tensor factorization (CGTF) that judiciously accounts for graph-related side information. The CGTF model has the potential to overcome practical challenges, such as missing slabs from the tensor and/or missing rows/columns from the correlation matrices. A novel alternating direction method of multipliers (ADMM) is also developed that recovers the nonnegative factors of CGTF. Our algorithm enjoys closed-form updates that result in reduced computational complexity and allow for convergence claims. A novel direction is further explored by employing the interpretable factors to detect graph communities having the tensor as side information. The resulting community detection approach is successful even when some links in the graphs are missing. Results with real data sets corroborate the merits of the proposed methods relative to state-of-the-art competing factorization techniques in providing recommendations and detecting communities.",
keywords = "Tensor-matrix factorization, community detection, graph data, recommender systems, tensor-graph imputation",
author = "Ioannidis, {Vassilis N.} and Zamzam, {Ahmed S.} and Giannakis, {Georgios B.} and Sidiropoulos, {Nicholas D.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1109/TKDE.2019.2941716",
language = "English (US)",
journal = "IEEE Transactions on Knowledge and Data Engineering",
issn = "1041-4347",
publisher = "IEEE Computer Society",

}

TY - JOUR

T1 - Coupled Graph and Tensor Factorization for Recommender Systems and Community Detection

AU - Ioannidis, Vassilis N.

AU - Zamzam, Ahmed S.

AU - Giannakis, Georgios B.

AU - Sidiropoulos, Nicholas D.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Joint analysis of data from multiple information repositories facilitates uncovering the underlying structure in heterogeneous datasets. Single and coupled matrix-tensor factorization (CMTF) has been widely used in this context for imputation-based recommendation from ratings, social network, and other user-item data. When this side information is in the form of item-item correlation matrices or graphs, existing CMTF algorithms may fall short. Alleviating current limitations, we introduce a novel model coined coupled graph-tensor factorization (CGTF) that judiciously accounts for graph-related side information. The CGTF model has the potential to overcome practical challenges, such as missing slabs from the tensor and/or missing rows/columns from the correlation matrices. A novel alternating direction method of multipliers (ADMM) is also developed that recovers the nonnegative factors of CGTF. Our algorithm enjoys closed-form updates that result in reduced computational complexity and allow for convergence claims. A novel direction is further explored by employing the interpretable factors to detect graph communities having the tensor as side information. The resulting community detection approach is successful even when some links in the graphs are missing. Results with real data sets corroborate the merits of the proposed methods relative to state-of-the-art competing factorization techniques in providing recommendations and detecting communities.

AB - Joint analysis of data from multiple information repositories facilitates uncovering the underlying structure in heterogeneous datasets. Single and coupled matrix-tensor factorization (CMTF) has been widely used in this context for imputation-based recommendation from ratings, social network, and other user-item data. When this side information is in the form of item-item correlation matrices or graphs, existing CMTF algorithms may fall short. Alleviating current limitations, we introduce a novel model coined coupled graph-tensor factorization (CGTF) that judiciously accounts for graph-related side information. The CGTF model has the potential to overcome practical challenges, such as missing slabs from the tensor and/or missing rows/columns from the correlation matrices. A novel alternating direction method of multipliers (ADMM) is also developed that recovers the nonnegative factors of CGTF. Our algorithm enjoys closed-form updates that result in reduced computational complexity and allow for convergence claims. A novel direction is further explored by employing the interpretable factors to detect graph communities having the tensor as side information. The resulting community detection approach is successful even when some links in the graphs are missing. Results with real data sets corroborate the merits of the proposed methods relative to state-of-the-art competing factorization techniques in providing recommendations and detecting communities.

KW - Tensor-matrix factorization

KW - community detection

KW - graph data

KW - recommender systems

KW - tensor-graph imputation

UR - http://www.scopus.com/inward/record.url?scp=85072530047&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85072530047&partnerID=8YFLogxK

U2 - 10.1109/TKDE.2019.2941716

DO - 10.1109/TKDE.2019.2941716

M3 - Article

AN - SCOPUS:85072530047

JO - IEEE Transactions on Knowledge and Data Engineering

JF - IEEE Transactions on Knowledge and Data Engineering

SN - 1041-4347

ER -