Unsupervised user behavior representation for fraud review detection with cold-start problem

Li, Q; Wu, Q; Zhu, C; Zhang, J; Zhao, W

Unsupervised user behavior representation for fraud review detection with cold-start problem

Li, Q

Wu, Q

Zhu, C

Zhang, J

Zhao, W

Permalink

Publication Type:: Conference Proceeding
Citation:: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2019, 11439 LNAI pp. 222 - 236
Issue Date:: 2019-01-01

Open Access

Copyright Clearance Process

Recently Added
In Progress
Open Access

This item is open access.

Adobe PDF

Download Accepted Manuscript versionAdobe PDF (1.04 MB)

View on publisher's site

View statistics

Full metadata record

Field	Value	Language
dc.contributor.author	Li, Q https://orcid.org/0000-0002-6149-5081	en_US
dc.contributor.author	Wu, Q https://orcid.org/0000-0001-5641-2483	en_US
dc.contributor.author	Zhu, C https://orcid.org/0000-0003-0495-1981	en_US
dc.contributor.author	Zhang, J https://orcid.org/0000-0002-7240-3541	en_US
dc.contributor.author	Zhao, W	en_US
dc.date.available	2020-05-25T19:19:28Z
dc.date.issued	2019-01-01	en_US
dc.identifier.citation	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2019, 11439 LNAI pp. 222 - 236	en_US
dc.identifier.isbn	9783030161477	en_US
dc.identifier.issn	0302-9743	en_US
dc.identifier.uri	http://hdl.handle.net/10453/132418
dc.description.abstract	© Springer Nature Switzerland AG 2019. Detecting fraud review is becoming extremely important in order to provide reliable information in cyberspace, in which, however, handling cold-start problem is a critical and urgent challenge since the case of cold-start fraud review rarely provides sufficient information for further assessing its authenticity. Existing work on detecting cold-start cases relies on the limited contents of the review posted by the user and a traditional classifier to make the decision. However, simply modeling review is not reliable since reviews can be easily manipulated. Also, it is hard to obtain high-quality labeled data for training the classifier. In this paper, we tackle cold-start problems by (1) using a user’s behavior representation rather than review contents to measure authenticity, which further (2) consider user social relations with other existing users when posting reviews. The method is completely (3) unsupervised. Comprehensive experiments on Yelp data sets demonstrate our method significantly outperforms the state-of-the-art methods.	en_US
dc.relation.ispartof	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)	en_US
dc.relation.isbasedon	10.1007/978-3-030-16148-4_18	en_US
dc.subject.classification	Artificial Intelligence & Image Processing	en_US
dc.title	Unsupervised user behavior representation for fraud review detection with cold-start problem	en_US
dc.type	Conference Proceeding
utslib.citation.volume	11439 LNAI	en_US
utslib.for	0803 Computer Software	en_US
pubs.embargo.period	Not known	en_US
pubs.organisational-group	/University of Technology Sydney
pubs.organisational-group	/University of Technology Sydney/Faculty of Engineering and Information Technology
pubs.organisational-group	/University of Technology Sydney/Faculty of Engineering and Information Technology/School of Electrical and Data Engineering
pubs.organisational-group	/University of Technology Sydney/Strength - GBDTC - Global Big Data Technologies
pubs.organisational-group	/University of Technology Sydney/Strength - INEXT - Innovation in IT Services and Applications
pubs.organisational-group	/University of Technology Sydney/Students
utslib.copyright.status	open_access
pubs.publication-status	Published	en_US
pubs.volume	11439 LNAI	en_US

Abstract:

© Springer Nature Switzerland AG 2019. Detecting fraud review is becoming extremely important in order to provide reliable information in cyberspace, in which, however, handling cold-start problem is a critical and urgent challenge since the case of cold-start fraud review rarely provides sufficient information for further assessing its authenticity. Existing work on detecting cold-start cases relies on the limited contents of the review posted by the user and a traditional classifier to make the decision. However, simply modeling review is not reliable since reviews can be easily manipulated. Also, it is hard to obtain high-quality labeled data for training the classifier. In this paper, we tackle cold-start problems by (1) using a user’s behavior representation rather than review contents to measure authenticity, which further (2) consider user social relations with other existing users when posting reviews. The method is completely (3) unsupervised. Comprehensive experiments on Yelp data sets demonstrate our method significantly outperforms the state-of-the-art methods.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/132418