Personalized difusions for top-n recommendation

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper introduces PerDif; a novel framework for learning personalized difusions over item-to-item graphs for top-n recommendation. PerDif learns the teleportation probabilities of a time-inhomogeneous random walk with restarts capturing a user-specifc underlying item exploration process. Such an approach can lead to signifcant improvements in recommendation accuracy, while also providing useful information about the users in the system. Per-user ftting can be performed in parallel and very efciently even in large-scale settings. A comprehensive set of experiments on real-world datasets demonstrate the scalability as well as the qualitative merits of the proposed framework. PerDif achieves high recommendation accuracy, outperforming state-of-the-art competing approaches-including several recently proposed methods relying on deep neural networks.

Original languageEnglish (US)
Title of host publicationRecSys 2019 - 13th ACM Conference on Recommender Systems
PublisherAssociation for Computing Machinery, Inc
Pages260-268
Number of pages9
ISBN (Electronic)9781450362436
DOIs
StatePublished - Sep 10 2019
Event13th ACM Conference on Recommender Systems, RecSys 2019 - Copenhagen, Denmark
Duration: Sep 16 2019Sep 20 2019

Publication series

NameRecSys 2019 - 13th ACM Conference on Recommender Systems

Conference

Conference13th ACM Conference on Recommender Systems, RecSys 2019
CountryDenmark
CityCopenhagen
Period9/16/199/20/19

Fingerprint

Scalability
Experiments
Deep neural networks

Keywords

  • Item Models
  • Random Walks
  • Top-N Recommendation

Cite this

Nikolakopoulos, A. N., Berberidis, D., Karypis, G., & Giannakis, G. B. (2019). Personalized difusions for top-n recommendation. In RecSys 2019 - 13th ACM Conference on Recommender Systems (pp. 260-268). (RecSys 2019 - 13th ACM Conference on Recommender Systems). Association for Computing Machinery, Inc. https://doi.org/10.1145/3298689.3346985

Personalized difusions for top-n recommendation. / Nikolakopoulos, Athanasios N.; Berberidis, Dimitris; Karypis, George; Giannakis, Georgios B.

RecSys 2019 - 13th ACM Conference on Recommender Systems. Association for Computing Machinery, Inc, 2019. p. 260-268 (RecSys 2019 - 13th ACM Conference on Recommender Systems).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Nikolakopoulos, AN, Berberidis, D, Karypis, G & Giannakis, GB 2019, Personalized difusions for top-n recommendation. in RecSys 2019 - 13th ACM Conference on Recommender Systems. RecSys 2019 - 13th ACM Conference on Recommender Systems, Association for Computing Machinery, Inc, pp. 260-268, 13th ACM Conference on Recommender Systems, RecSys 2019, Copenhagen, Denmark, 9/16/19. https://doi.org/10.1145/3298689.3346985
Nikolakopoulos AN, Berberidis D, Karypis G, Giannakis GB. Personalized difusions for top-n recommendation. In RecSys 2019 - 13th ACM Conference on Recommender Systems. Association for Computing Machinery, Inc. 2019. p. 260-268. (RecSys 2019 - 13th ACM Conference on Recommender Systems). https://doi.org/10.1145/3298689.3346985
Nikolakopoulos, Athanasios N. ; Berberidis, Dimitris ; Karypis, George ; Giannakis, Georgios B. / Personalized difusions for top-n recommendation. RecSys 2019 - 13th ACM Conference on Recommender Systems. Association for Computing Machinery, Inc, 2019. pp. 260-268 (RecSys 2019 - 13th ACM Conference on Recommender Systems).
@inproceedings{e0a18a5c77624130b543dd81942b6db0,
title = "Personalized difusions for top-n recommendation",
abstract = "This paper introduces PerDif; a novel framework for learning personalized difusions over item-to-item graphs for top-n recommendation. PerDif learns the teleportation probabilities of a time-inhomogeneous random walk with restarts capturing a user-specifc underlying item exploration process. Such an approach can lead to signifcant improvements in recommendation accuracy, while also providing useful information about the users in the system. Per-user ftting can be performed in parallel and very efciently even in large-scale settings. A comprehensive set of experiments on real-world datasets demonstrate the scalability as well as the qualitative merits of the proposed framework. PerDif achieves high recommendation accuracy, outperforming state-of-the-art competing approaches-including several recently proposed methods relying on deep neural networks.",
keywords = "Item Models, Random Walks, Top-N Recommendation",
author = "Nikolakopoulos, {Athanasios N.} and Dimitris Berberidis and George Karypis and Giannakis, {Georgios B.}",
year = "2019",
month = "9",
day = "10",
doi = "10.1145/3298689.3346985",
language = "English (US)",
series = "RecSys 2019 - 13th ACM Conference on Recommender Systems",
publisher = "Association for Computing Machinery, Inc",
pages = "260--268",
booktitle = "RecSys 2019 - 13th ACM Conference on Recommender Systems",

}

TY - GEN

T1 - Personalized difusions for top-n recommendation

AU - Nikolakopoulos, Athanasios N.

AU - Berberidis, Dimitris

AU - Karypis, George

AU - Giannakis, Georgios B.

PY - 2019/9/10

Y1 - 2019/9/10

N2 - This paper introduces PerDif; a novel framework for learning personalized difusions over item-to-item graphs for top-n recommendation. PerDif learns the teleportation probabilities of a time-inhomogeneous random walk with restarts capturing a user-specifc underlying item exploration process. Such an approach can lead to signifcant improvements in recommendation accuracy, while also providing useful information about the users in the system. Per-user ftting can be performed in parallel and very efciently even in large-scale settings. A comprehensive set of experiments on real-world datasets demonstrate the scalability as well as the qualitative merits of the proposed framework. PerDif achieves high recommendation accuracy, outperforming state-of-the-art competing approaches-including several recently proposed methods relying on deep neural networks.

AB - This paper introduces PerDif; a novel framework for learning personalized difusions over item-to-item graphs for top-n recommendation. PerDif learns the teleportation probabilities of a time-inhomogeneous random walk with restarts capturing a user-specifc underlying item exploration process. Such an approach can lead to signifcant improvements in recommendation accuracy, while also providing useful information about the users in the system. Per-user ftting can be performed in parallel and very efciently even in large-scale settings. A comprehensive set of experiments on real-world datasets demonstrate the scalability as well as the qualitative merits of the proposed framework. PerDif achieves high recommendation accuracy, outperforming state-of-the-art competing approaches-including several recently proposed methods relying on deep neural networks.

KW - Item Models

KW - Random Walks

KW - Top-N Recommendation

UR - http://www.scopus.com/inward/record.url?scp=85073378484&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85073378484&partnerID=8YFLogxK

U2 - 10.1145/3298689.3346985

DO - 10.1145/3298689.3346985

M3 - Conference contribution

AN - SCOPUS:85073378484

T3 - RecSys 2019 - 13th ACM Conference on Recommender Systems

SP - 260

EP - 268

BT - RecSys 2019 - 13th ACM Conference on Recommender Systems

PB - Association for Computing Machinery, Inc

ER -