MSTAN: A Multi-view Spatio-Temporal Aggregation Network Learning Irregular Interval User Activities for Fraud Detection

Wenbo Zhang; Shuo Zhang; Xingbang Hu; Hejiao Huang

doi:10.1007/978-981-97-2262-4_31

MSTAN: A Multi-view Spatio-Temporal Aggregation Network Learning Irregular Interval User Activities for Fraud Detection

Wenbo Zhang
, Shuo Zhang
, Xingbang Hu
, Hejiao Huang^*

^*Corresponding author for this work

School of Computer Science and Technology, Harbin Institute of Technology
Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies

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

1 Scopus citations

Abstract

Discovering fraud patterns from numerous user activities is crucial for fraud detection. However, three factors make this task quite challenging: Firstly, previous research usually utilize just one of the two forms of user activity, namely sequential behavior and interaction relationship, leaving much information unused. Additionally, nearly all works merely study on a single view of user activities, but fraud patterns often span across multiple views. Moreover, most existing models can only handle regular time intervals, while in reality, user activities occur with irregular time intervals. To effectively discover fraud patterns from user activities, this paper proposes MSTAN (Multi-view Spatio-Temporal Aggregation Network) for fraud detection. It addresses the above problems through three phases: (1) In short-term aggregation, SIFB (Sequential behavior and Interaction relationship Fusion Block) is employed to integrate sequential behavior and interaction relationship. (2) In view aggregation, 2-dimensional multi-view user activity embedding is obtained for simultaneously mining multiple views. (3) In long-term aggregation CTLSTM (Convolutional Time LSTM) is designed to deal with irregular time intervals. Experiments on two real world datasets demonstrate that our model outperforms the comparison methods.

Original language	English
Title of host publication	Advances in Knowledge Discovery and Data Mining - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024, Proceedings
Editors	De-Nian Yang, Xing Xie, Vincent S. Tseng, Jian Pei, Jen-Wei Huang, Jerry Chun-Wei Lin
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	389-401
Number of pages	13
ISBN (Print)	9789819722648
DOIs	https://doi.org/10.1007/978-981-97-2262-4_31
State	Published - 2024
Externally published	Yes
Event	28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024 - Taipei, Taiwan, Province of China Duration: 7 May 2024 → 10 May 2024

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14649 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024
Country/Territory	Taiwan, Province of China
City	Taipei
Period	7/05/24 → 10/05/24

Keywords

Attention mechanism
Fraud detection
Neural networks
Spatio-temporal aggregation

Access to Document

10.1007/978-981-97-2262-4_31

Cite this

Zhang, W., Zhang, S., Hu, X., & Huang, H. (2024). MSTAN: A Multi-view Spatio-Temporal Aggregation Network Learning Irregular Interval User Activities for Fraud Detection. In D.-N. Yang, X. Xie, V. S. Tseng, J. Pei, J.-W. Huang, & J. C.-W. Lin (Eds.), Advances in Knowledge Discovery and Data Mining - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024, Proceedings (pp. 389-401). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14649 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-981-97-2262-4_31

Zhang, Wenbo ; Zhang, Shuo ; Hu, Xingbang et al. / MSTAN : A Multi-view Spatio-Temporal Aggregation Network Learning Irregular Interval User Activities for Fraud Detection. Advances in Knowledge Discovery and Data Mining - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024, Proceedings. editor / De-Nian Yang ; Xing Xie ; Vincent S. Tseng ; Jian Pei ; Jen-Wei Huang ; Jerry Chun-Wei Lin. Springer Science and Business Media Deutschland GmbH, 2024. pp. 389-401 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{9740f164326c421e970f262762856465,

title = "MSTAN: A Multi-view Spatio-Temporal Aggregation Network Learning Irregular Interval User Activities for Fraud Detection",

abstract = "Discovering fraud patterns from numerous user activities is crucial for fraud detection. However, three factors make this task quite challenging: Firstly, previous research usually utilize just one of the two forms of user activity, namely sequential behavior and interaction relationship, leaving much information unused. Additionally, nearly all works merely study on a single view of user activities, but fraud patterns often span across multiple views. Moreover, most existing models can only handle regular time intervals, while in reality, user activities occur with irregular time intervals. To effectively discover fraud patterns from user activities, this paper proposes MSTAN (Multi-view Spatio-Temporal Aggregation Network) for fraud detection. It addresses the above problems through three phases: (1) In short-term aggregation, SIFB (Sequential behavior and Interaction relationship Fusion Block) is employed to integrate sequential behavior and interaction relationship. (2) In view aggregation, 2-dimensional multi-view user activity embedding is obtained for simultaneously mining multiple views. (3) In long-term aggregation CTLSTM (Convolutional Time LSTM) is designed to deal with irregular time intervals. Experiments on two real world datasets demonstrate that our model outperforms the comparison methods.",

keywords = "Attention mechanism, Fraud detection, Neural networks, Spatio-temporal aggregation",

author = "Wenbo Zhang and Shuo Zhang and Xingbang Hu and Hejiao Huang",

note = "Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.; 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024 ; Conference date: 07-05-2024 Through 10-05-2024",

year = "2024",

doi = "10.1007/978-981-97-2262-4\_31",

language = "英语",

isbn = "9789819722648",

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 = "389--401",

editor = "De-Nian Yang and Xing Xie and Tseng, \{Vincent S.\} and Jian Pei and Jen-Wei Huang and Lin, \{Jerry Chun-Wei\}",

booktitle = "Advances in Knowledge Discovery and Data Mining - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024, Proceedings",

address = "德国",

}

Zhang, W, Zhang, S, Hu, X & Huang, H 2024, MSTAN: A Multi-view Spatio-Temporal Aggregation Network Learning Irregular Interval User Activities for Fraud Detection. in D-N Yang, X Xie, VS Tseng, J Pei, J-W Huang & JC-W Lin (eds), Advances in Knowledge Discovery and Data Mining - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14649 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 389-401, 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024, Taipei, Taiwan, Province of China, 7/05/24. https://doi.org/10.1007/978-981-97-2262-4_31

MSTAN: A Multi-view Spatio-Temporal Aggregation Network Learning Irregular Interval User Activities for Fraud Detection. / Zhang, Wenbo; Zhang, Shuo; Hu, Xingbang et al.
Advances in Knowledge Discovery and Data Mining - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024, Proceedings. ed. / De-Nian Yang; Xing Xie; Vincent S. Tseng; Jian Pei; Jen-Wei Huang; Jerry Chun-Wei Lin. Springer Science and Business Media Deutschland GmbH, 2024. p. 389-401 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14649 LNAI).

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

TY - GEN

T1 - MSTAN

T2 - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024

AU - Zhang, Wenbo

AU - Zhang, Shuo

AU - Hu, Xingbang

AU - Huang, Hejiao

N1 - Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

PY - 2024

Y1 - 2024

N2 - Discovering fraud patterns from numerous user activities is crucial for fraud detection. However, three factors make this task quite challenging: Firstly, previous research usually utilize just one of the two forms of user activity, namely sequential behavior and interaction relationship, leaving much information unused. Additionally, nearly all works merely study on a single view of user activities, but fraud patterns often span across multiple views. Moreover, most existing models can only handle regular time intervals, while in reality, user activities occur with irregular time intervals. To effectively discover fraud patterns from user activities, this paper proposes MSTAN (Multi-view Spatio-Temporal Aggregation Network) for fraud detection. It addresses the above problems through three phases: (1) In short-term aggregation, SIFB (Sequential behavior and Interaction relationship Fusion Block) is employed to integrate sequential behavior and interaction relationship. (2) In view aggregation, 2-dimensional multi-view user activity embedding is obtained for simultaneously mining multiple views. (3) In long-term aggregation CTLSTM (Convolutional Time LSTM) is designed to deal with irregular time intervals. Experiments on two real world datasets demonstrate that our model outperforms the comparison methods.

AB - Discovering fraud patterns from numerous user activities is crucial for fraud detection. However, three factors make this task quite challenging: Firstly, previous research usually utilize just one of the two forms of user activity, namely sequential behavior and interaction relationship, leaving much information unused. Additionally, nearly all works merely study on a single view of user activities, but fraud patterns often span across multiple views. Moreover, most existing models can only handle regular time intervals, while in reality, user activities occur with irregular time intervals. To effectively discover fraud patterns from user activities, this paper proposes MSTAN (Multi-view Spatio-Temporal Aggregation Network) for fraud detection. It addresses the above problems through three phases: (1) In short-term aggregation, SIFB (Sequential behavior and Interaction relationship Fusion Block) is employed to integrate sequential behavior and interaction relationship. (2) In view aggregation, 2-dimensional multi-view user activity embedding is obtained for simultaneously mining multiple views. (3) In long-term aggregation CTLSTM (Convolutional Time LSTM) is designed to deal with irregular time intervals. Experiments on two real world datasets demonstrate that our model outperforms the comparison methods.

KW - Attention mechanism

KW - Fraud detection

KW - Neural networks

KW - Spatio-temporal aggregation

UR - https://www.scopus.com/pages/publications/85192854002

U2 - 10.1007/978-981-97-2262-4_31

DO - 10.1007/978-981-97-2262-4_31

M3 - 会议稿件

AN - SCOPUS:85192854002

SN - 9789819722648

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

SP - 389

EP - 401

BT - Advances in Knowledge Discovery and Data Mining - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024, Proceedings

A2 - Yang, De-Nian

A2 - Xie, Xing

A2 - Tseng, Vincent S.

A2 - Pei, Jian

A2 - Huang, Jen-Wei

A2 - Lin, Jerry Chun-Wei

PB - Springer Science and Business Media Deutschland GmbH

Y2 - 7 May 2024 through 10 May 2024

ER -

Zhang W, Zhang S, Hu X, Huang H. MSTAN: A Multi-view Spatio-Temporal Aggregation Network Learning Irregular Interval User Activities for Fraud Detection. In Yang DN, Xie X, Tseng VS, Pei J, Huang JW, Lin JCW, editors, Advances in Knowledge Discovery and Data Mining - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024, Proceedings. Springer Science and Business Media Deutschland GmbH. 2024. p. 389-401. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-981-97-2262-4_31