Modeling user hidden navigational behavior for Web recommendation

Xu, G; Li, L; Zhang, Y; Yi, X; Kitsuregawa, M

Modeling user hidden navigational behavior for Web recommendation

Xu, G

Li, L Zhang, Y Yi, X Kitsuregawa, M

Permalink

Publication Type:: Journal Article
Citation:: Web Intelligence and Agent Systems, 2011, 9 (3), pp. 239 - 255
Issue Date:: 2011-08-15

Closed Access

	Filename	Description	Size
	2013002779OK.pdf		594.95 kB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Closed Access

This item is closed access and not available.

Full metadata record

Field	Value	Language
dc.contributor.author	Xu, G https://orcid.org/0000-0003-4493-6663	en_US
dc.contributor.author	Li, L	en_US
dc.contributor.author	Zhang, Y	en_US
dc.contributor.author	Yi, X	en_US
dc.contributor.author	Kitsuregawa, M	en_US
dc.date.issued	2011-08-15	en_US
dc.identifier.citation	Web Intelligence and Agent Systems, 2011, 9 (3), pp. 239 - 255	en_US
dc.identifier.issn	1570-1263	en_US
dc.identifier.uri	http://hdl.handle.net/10453/29139
dc.identifier.uri	http://hdl.handle.net/10453/33757
dc.description.abstract	Web users exhibit a variety of navigational interests through clicking a sequence of Web pages. Analyses of Web usage data will lead to discovering Web user access patterns, and in turn, facilitating users to locate more preferable Web contents via collaborative recommendation techniques. In the context of Web usage mining, Latent Semantic Analysis (LSA) based on probability inference provides a promising approach to capture not only user hidden navigational patterns, but also the associations between users, pages and hidden navigational patterns. The discovered user access patterns could be used as a usage reference base for identifying the new user's access preferences and making usage-based collaborative Web recommendations. In this paper, we propose a novel usage-based Web recommendation framework, in which Latent Dirichlet Allocation (LDA) is employed to learn the underlying task space from the training Web log data and infer the task distribution for a target user via task inference. The main advantages of the adapted LDA model are its capabilities of efficiently learning the semantic usage information from the Web log data and effectively inferring the access preference of the target user even with a few Web clicks that might be unseen in the training data. In this paper, we also investigate the determination of an optimizing task number, which is another important problem commonly encountered in latent semantic analysis. Experiments conducted on a real Web log dataset show that this approach can achieve better recommendation performance in comparison to other existing techniques. And the discovered task-simplex expression can also provide a better interpretation for Web designers or users to better understand the user navigational preference. © 2011 - IOS Press and the authors. All rights reserved.	en_US
dc.relation.ispartof	Web Intelligence and Agent Systems	en_US
dc.relation.isbasedon	10.3233/WIA-2011-0217	en_US
dc.subject.classification	Artificial Intelligence & Image Processing	en_US
dc.title	Modeling user hidden navigational behavior for Web recommendation	en_US
dc.type	Journal Article
utslib.citation.volume	3	en_US
utslib.citation.volume	9	en_US
utslib.for	0801 Artificial Intelligence and Image Processing	en_US
utslib.for	0899 Other Information and Computing Sciences	en_US
utslib.for	1702 Cognitive Sciences	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 Computer Science
pubs.organisational-group	/University of Technology Sydney/Strength - AAI - Advanced Analytics Institute Research Centre
utslib.copyright.status	closed_access
pubs.issue	3	en_US
pubs.publication-status	Published	en_US
pubs.volume	9	en_US

Abstract:

Web users exhibit a variety of navigational interests through clicking a sequence of Web pages. Analyses of Web usage data will lead to discovering Web user access patterns, and in turn, facilitating users to locate more preferable Web contents via collaborative recommendation techniques. In the context of Web usage mining, Latent Semantic Analysis (LSA) based on probability inference provides a promising approach to capture not only user hidden navigational patterns, but also the associations between users, pages and hidden navigational patterns. The discovered user access patterns could be used as a usage reference base for identifying the new user's access preferences and making usage-based collaborative Web recommendations. In this paper, we propose a novel usage-based Web recommendation framework, in which Latent Dirichlet Allocation (LDA) is employed to learn the underlying task space from the training Web log data and infer the task distribution for a target user via task inference. The main advantages of the adapted LDA model are its capabilities of efficiently learning the semantic usage information from the Web log data and effectively inferring the access preference of the target user even with a few Web clicks that might be unseen in the training data. In this paper, we also investigate the determination of an optimizing task number, which is another important problem commonly encountered in latent semantic analysis. Experiments conducted on a real Web log dataset show that this approach can achieve better recommendation performance in comparison to other existing techniques. And the discovered task-simplex expression can also provide a better interpretation for Web designers or users to better understand the user navigational preference. © 2011 - IOS Press and the authors. All rights reserved.

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

http://hdl.handle.net/10453/29139