Empirical Comparison of Graph Embeddings for Trust-Based Collaborative Filtering

Tomislav Duricic; Hussain Hussain; Emanuel Lacic; Dominik Kowald; Denis Helic; Elisabeth Lex

doi:10.1007/978-3-030-59491-6_17

Empirical Comparison of Graph Embeddings for Trust-Based Collaborative Filtering

Tomislav Duricic^*, Hussain Hussain, Emanuel Lacic, Dominik Kowald, Denis Helic, Elisabeth Lex

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

Abstract

In this work, we study the utility of graph embeddings to generate latent user representations for trust-based collaborative filtering. In a cold-start setting, on three publicly available datasets, we evaluate approaches from four method families: (i) factorization-based, (ii) random walk-based, (iii) deep learning-based, and (iv) the Large-scale Information Network Embedding (LINE) approach. We find that across the four families, random-walk-based approaches consistently achieve the best accuracy. Besides, they result in highly novel and diverse recommendations. Furthermore, our results show that the use of graph embeddings in trust-based collaborative filtering significantly improves user coverage.

Original language	English
Title of host publication	Foundations of Intelligent Systems - 25th International Symposium, ISMIS 2020, Proceedings
Editors	Denis Helic, Martin Stettinger, Alexander Felfernig, Gerhard Leitner, Zbigniew W. Ras
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	181-191
Number of pages	11
ISBN (Print)	9783030594909
DOIs	https://doi.org/10.1007/978-3-030-59491-6_17
Publication status	Published - 1 Jan 2020
Event	25th International Symposium on Methodologies for Intelligent Systems - TU Graz, Virtuell, Austria Duration: 23 Sept 2020 → 25 Sept 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12117 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	25th International Symposium on Methodologies for Intelligent Systems
Abbreviated title	ISMIS 2020
Country/Territory	Austria
City	Virtuell
Period	23/09/20 → 25/09/20

Keywords

Cold-start
Empirical study
Graph embeddings
Recommender systems
Trust networks

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-030-59491-6_17

Cite this

Duricic, T., Hussain, H., Lacic, E., Kowald, D., Helic, D., & Lex, E. (2020). Empirical Comparison of Graph Embeddings for Trust-Based Collaborative Filtering. In D. Helic, M. Stettinger, A. Felfernig, G. Leitner, & Z. W. Ras (Eds.), Foundations of Intelligent Systems - 25th International Symposium, ISMIS 2020, Proceedings (pp. 181-191). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12117 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-59491-6_17

Empirical Comparison of Graph Embeddings for Trust-Based Collaborative Filtering. / Duricic, Tomislav; Hussain, Hussain; Lacic, Emanuel et al.
Foundations of Intelligent Systems - 25th International Symposium, ISMIS 2020, Proceedings. ed. / Denis Helic; Martin Stettinger; Alexander Felfernig; Gerhard Leitner; Zbigniew W. Ras. Springer Science and Business Media Deutschland GmbH, 2020. p. 181-191 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12117 LNAI).

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

Duricic, T, Hussain, H, Lacic, E, Kowald, D , Helic, D & Lex, E 2020, Empirical Comparison of Graph Embeddings for Trust-Based Collaborative Filtering. in D Helic, M Stettinger, A Felfernig, G Leitner & ZW Ras (eds), Foundations of Intelligent Systems - 25th International Symposium, ISMIS 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12117 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 181-191, 25th International Symposium on Methodologies for Intelligent Systems, Virtuell, Austria, 23/09/20. https://doi.org/10.1007/978-3-030-59491-6_17

Duricic T, Hussain H, Lacic E, Kowald D , Helic D , Lex E. Empirical Comparison of Graph Embeddings for Trust-Based Collaborative Filtering. In Helic D, Stettinger M, Felfernig A, Leitner G, Ras ZW, editors, Foundations of Intelligent Systems - 25th International Symposium, ISMIS 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2020. p. 181-191. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-59491-6_17

Duricic, Tomislav ; Hussain, Hussain ; Lacic, Emanuel et al. / Empirical Comparison of Graph Embeddings for Trust-Based Collaborative Filtering. Foundations of Intelligent Systems - 25th International Symposium, ISMIS 2020, Proceedings. editor / Denis Helic ; Martin Stettinger ; Alexander Felfernig ; Gerhard Leitner ; Zbigniew W. Ras. Springer Science and Business Media Deutschland GmbH, 2020. pp. 181-191 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{f4cdc2acaecc4ff3ab0113c067ee1cca,

title = "Empirical Comparison of Graph Embeddings for Trust-Based Collaborative Filtering",

abstract = "In this work, we study the utility of graph embeddings to generate latent user representations for trust-based collaborative filtering. In a cold-start setting, on three publicly available datasets, we evaluate approaches from four method families: (i) factorization-based, (ii) random walk-based, (iii) deep learning-based, and (iv) the Large-scale Information Network Embedding (LINE) approach. We find that across the four families, random-walk-based approaches consistently achieve the best accuracy. Besides, they result in highly novel and diverse recommendations. Furthermore, our results show that the use of graph embeddings in trust-based collaborative filtering significantly improves user coverage.",

keywords = "Cold-start, Empirical study, Graph embeddings, Recommender systems, Trust networks",

author = "Tomislav Duricic and Hussain Hussain and Emanuel Lacic and Dominik Kowald and Denis Helic and Elisabeth Lex",

year = "2020",

month = jan,

day = "1",

doi = "10.1007/978-3-030-59491-6_17",

language = "English",

isbn = "9783030594909",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "181--191",

editor = "Denis Helic and Martin Stettinger and Alexander Felfernig and Gerhard Leitner and Ras, {Zbigniew W.}",

booktitle = "Foundations of Intelligent Systems - 25th International Symposium, ISMIS 2020, Proceedings",

address = "Germany",

note = "25th International Symposium on Methodologies for Intelligent Systems, ISMIS 2020 ; Conference date: 23-09-2020 Through 25-09-2020",

}

TY - GEN

T1 - Empirical Comparison of Graph Embeddings for Trust-Based Collaborative Filtering

AU - Duricic, Tomislav

AU - Hussain, Hussain

AU - Lacic, Emanuel

AU - Kowald, Dominik

AU - Helic, Denis

AU - Lex, Elisabeth

PY - 2020/1/1

Y1 - 2020/1/1

N2 - In this work, we study the utility of graph embeddings to generate latent user representations for trust-based collaborative filtering. In a cold-start setting, on three publicly available datasets, we evaluate approaches from four method families: (i) factorization-based, (ii) random walk-based, (iii) deep learning-based, and (iv) the Large-scale Information Network Embedding (LINE) approach. We find that across the four families, random-walk-based approaches consistently achieve the best accuracy. Besides, they result in highly novel and diverse recommendations. Furthermore, our results show that the use of graph embeddings in trust-based collaborative filtering significantly improves user coverage.

AB - In this work, we study the utility of graph embeddings to generate latent user representations for trust-based collaborative filtering. In a cold-start setting, on three publicly available datasets, we evaluate approaches from four method families: (i) factorization-based, (ii) random walk-based, (iii) deep learning-based, and (iv) the Large-scale Information Network Embedding (LINE) approach. We find that across the four families, random-walk-based approaches consistently achieve the best accuracy. Besides, they result in highly novel and diverse recommendations. Furthermore, our results show that the use of graph embeddings in trust-based collaborative filtering significantly improves user coverage.

KW - Cold-start

KW - Empirical study

KW - Graph embeddings

KW - Recommender systems

KW - Trust networks

UR - http://www.scopus.com/inward/record.url?scp=85092113020&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-59491-6_17

DO - 10.1007/978-3-030-59491-6_17

M3 - Conference paper

AN - SCOPUS:85092113020

SN - 9783030594909

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 181

EP - 191

BT - Foundations of Intelligent Systems - 25th International Symposium, ISMIS 2020, Proceedings

A2 - Helic, Denis

A2 - Stettinger, Martin

A2 - Felfernig, Alexander

A2 - Leitner, Gerhard

A2 - Ras, Zbigniew W.

PB - Springer Science and Business Media Deutschland GmbH

T2 - 25th International Symposium on Methodologies for Intelligent Systems

Y2 - 23 September 2020 through 25 September 2020

ER -

Empirical Comparison of Graph Embeddings for Trust-Based Collaborative Filtering

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this