Temporal Heterogeneous Interaction Graph Embedding for Next-Item Recommendation

Yugang Ji; Ming Yang Yin; Yuan Fang; Hongxia Yang; Xiangwei Wang; Tianrui Jia; Chuan Shi

doi:10.1007/978-3-030-67664-3_19

Temporal Heterogeneous Interaction Graph Embedding for Next-Item Recommendation

Yugang Ji
, Ming Yang Yin
, Yuan Fang
, Hongxia Yang
, Xiangwei Wang
, Tianrui Jia
, Chuan Shi

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

12 Citations (Scopus)

Abstract

In the scenario of next-item recommendation, previous methods attempt to model user preferences by capturing the evolution of sequential interactions. However, their sequential expression is often limited, without modeling complex dynamics that short-term demands can often be influenced by long-term habits. Moreover, few of them take into account the heterogeneous types of interaction between users and items. In this paper, we model such complex data as a Temporal Heterogeneous Interaction Graph (THIG) and learn both user and item embeddings on THIGs to address next-item recommendation. The main challenges involve two aspects: the complex dynamics and rich heterogeneity of interactions. We propose THIG Embedding (THIGE) which models the complex dynamics so that evolving short-term demands are guided by long-term historical habits, and leverages the rich heterogeneity to express the latent relevance of different-typed preferences. Extensive experiments on real-world datasets demonstrate that THIGE consistently outperforms the state-of-the-art methods.

Original language	English
Title of host publication	Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2020, Proceedings
Editors	Frank Hutter, Kristian Kersting, Jefrey Lijffijt, Isabel Valera
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	314-329
Number of pages	16
ISBN (Print)	9783030676636
DOIs	https://doi.org/10.1007/978-3-030-67664-3_19
Publication status	Published - Feb 2021
Externally published	Yes
Event	European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2020 - Virtual, Online Duration: 14 Sept 2020 → 18 Sept 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12459 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2020
City	Virtual, Online
Period	14/09/20 → 18/09/20

Keywords

Long-term habits
Next-item recommendation
Short-term demands
Temporal heterogeneous interaction graph

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

More information

10.1007/978-3-030-67664-3_19

Cite this

Ji, Y., Yin, M. Y., Fang, Y., Yang, H., Wang, X., Jia, T., & Shi, C. (2021). Temporal Heterogeneous Interaction Graph Embedding for Next-Item Recommendation. In F. Hutter, K. Kersting, J. Lijffijt, & I. Valera (Eds.), Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2020, Proceedings (pp. 314-329). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12459 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-67664-3_19

Ji, Yugang ; Yin, Ming Yang ; Fang, Yuan et al. / Temporal Heterogeneous Interaction Graph Embedding for Next-Item Recommendation. Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2020, Proceedings. editor / Frank Hutter ; Kristian Kersting ; Jefrey Lijffijt ; Isabel Valera. Springer Science and Business Media Deutschland GmbH, 2021. pp. 314-329 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{ef8d00a4fdb248a7b04fc5efb2edc768,

title = "Temporal Heterogeneous Interaction Graph Embedding for Next-Item Recommendation",

abstract = "In the scenario of next-item recommendation, previous methods attempt to model user preferences by capturing the evolution of sequential interactions. However, their sequential expression is often limited, without modeling complex dynamics that short-term demands can often be influenced by long-term habits. Moreover, few of them take into account the heterogeneous types of interaction between users and items. In this paper, we model such complex data as a Temporal Heterogeneous Interaction Graph (THIG) and learn both user and item embeddings on THIGs to address next-item recommendation. The main challenges involve two aspects: the complex dynamics and rich heterogeneity of interactions. We propose THIG Embedding (THIGE) which models the complex dynamics so that evolving short-term demands are guided by long-term historical habits, and leverages the rich heterogeneity to express the latent relevance of different-typed preferences. Extensive experiments on real-world datasets demonstrate that THIGE consistently outperforms the state-of-the-art methods.",

keywords = "Long-term habits, Next-item recommendation, Short-term demands, Temporal heterogeneous interaction graph",

author = "Yugang Ji and Yin, \{Ming Yang\} and Yuan Fang and Hongxia Yang and Xiangwei Wang and Tianrui Jia and Chuan Shi",

note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2020 ; Conference date: 14-09-2020 Through 18-09-2020",

year = "2021",

month = feb,

doi = "10.1007/978-3-030-67664-3\_19",

language = "English",

isbn = "9783030676636",

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 = "314--329",

editor = "Frank Hutter and Kristian Kersting and Jefrey Lijffijt and Isabel Valera",

booktitle = "Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2020, Proceedings",

address = "Germany",

}

Ji, Y, Yin, MY, Fang, Y, Yang, H, Wang, X, Jia, T & Shi, C 2021, Temporal Heterogeneous Interaction Graph Embedding for Next-Item Recommendation. in F Hutter, K Kersting, J Lijffijt & I Valera (eds), Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12459 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 314-329, European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2020, Virtual, Online, 14/09/20. https://doi.org/10.1007/978-3-030-67664-3_19

Temporal Heterogeneous Interaction Graph Embedding for Next-Item Recommendation. / Ji, Yugang; Yin, Ming Yang; Fang, Yuan et al.
Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2020, Proceedings. ed. / Frank Hutter; Kristian Kersting; Jefrey Lijffijt; Isabel Valera. Springer Science and Business Media Deutschland GmbH, 2021. p. 314-329 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12459 LNAI).

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

TY - GEN

T1 - Temporal Heterogeneous Interaction Graph Embedding for Next-Item Recommendation

AU - Ji, Yugang

AU - Yin, Ming Yang

AU - Fang, Yuan

AU - Yang, Hongxia

AU - Wang, Xiangwei

AU - Jia, Tianrui

AU - Shi, Chuan

PY - 2021/2

Y1 - 2021/2

N2 - In the scenario of next-item recommendation, previous methods attempt to model user preferences by capturing the evolution of sequential interactions. However, their sequential expression is often limited, without modeling complex dynamics that short-term demands can often be influenced by long-term habits. Moreover, few of them take into account the heterogeneous types of interaction between users and items. In this paper, we model such complex data as a Temporal Heterogeneous Interaction Graph (THIG) and learn both user and item embeddings on THIGs to address next-item recommendation. The main challenges involve two aspects: the complex dynamics and rich heterogeneity of interactions. We propose THIG Embedding (THIGE) which models the complex dynamics so that evolving short-term demands are guided by long-term historical habits, and leverages the rich heterogeneity to express the latent relevance of different-typed preferences. Extensive experiments on real-world datasets demonstrate that THIGE consistently outperforms the state-of-the-art methods.

AB - In the scenario of next-item recommendation, previous methods attempt to model user preferences by capturing the evolution of sequential interactions. However, their sequential expression is often limited, without modeling complex dynamics that short-term demands can often be influenced by long-term habits. Moreover, few of them take into account the heterogeneous types of interaction between users and items. In this paper, we model such complex data as a Temporal Heterogeneous Interaction Graph (THIG) and learn both user and item embeddings on THIGs to address next-item recommendation. The main challenges involve two aspects: the complex dynamics and rich heterogeneity of interactions. We propose THIG Embedding (THIGE) which models the complex dynamics so that evolving short-term demands are guided by long-term historical habits, and leverages the rich heterogeneity to express the latent relevance of different-typed preferences. Extensive experiments on real-world datasets demonstrate that THIGE consistently outperforms the state-of-the-art methods.

KW - Long-term habits

KW - Next-item recommendation

KW - Short-term demands

KW - Temporal heterogeneous interaction graph

UR - https://www.scopus.com/pages/publications/85103296837

U2 - 10.1007/978-3-030-67664-3_19

DO - 10.1007/978-3-030-67664-3_19

M3 - Conference article published in proceeding or book

AN - SCOPUS:85103296837

SN - 9783030676636

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

SP - 314

EP - 329

BT - Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2020, Proceedings

A2 - Hutter, Frank

A2 - Kersting, Kristian

A2 - Lijffijt, Jefrey

A2 - Valera, Isabel

PB - Springer Science and Business Media Deutschland GmbH

T2 - European Conference on Machine Learning and Knowledge Discovery in Databases, ECML PKDD 2020

Y2 - 14 September 2020 through 18 September 2020

ER -

Ji Y, Yin MY, Fang Y, Yang H, Wang X, Jia T et al. Temporal Heterogeneous Interaction Graph Embedding for Next-Item Recommendation. In Hutter F, Kersting K, Lijffijt J, Valera I, editors, Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2021. p. 314-329. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-67664-3_19

Temporal Heterogeneous Interaction Graph Embedding for Next-Item Recommendation

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

More information

Other files and links

Fingerprint

Cite this