Guandong Xu - Leveraging user review and numeric rating order to capture user preference in location-based social networks

14:00
Thursday
3
Sep
2015
Organized by: 
Eric Gaussier
Speaker: 
Guandong Xu

Detailed information: 

Dr Guandong Xu is a senior lecturer in the Advanced Analytics Institute at University of Technology Sydney. He received MSc and BSc degree in Computer Science and Engineering from Zhejiang University, China. He gained PhD degree in Computer Science from Victoria University. After that he took various positions, e.g., Postdoctoral research fellow and Vice-Chancellor Postdoctoral Fellow in the Centre for Applied Informatics at Victoria University, Australia, and Research Assistant Professor in Department of Computer Science at Aalborg University, Denmark. He is an Endeavour Postdoctoral Research Fellow in the University of Tokyo in 2008.

Conference venue :

(See file "howtogotoama")
Room Turing
Laboratoire LIG - Bâtiment Centre Equation 4
Allée de la Palestine à Gières

38610 GIERES

Tel – 04 76 51 46 24
 
Abstract: 

With the pervasive use of mobile devices, Location Based Social Networks (LBSNs) have emerged and become popular in past years. These LBSNs, allowing their users to share personal experiences and opinions on visited merchants, have very rich and useful information which enables a new breed of location-based services, namely, Merchant Recommendation. Existing techniques for merchant recommendation simply rely on rating data and treat each merchant as an item, and apply conventional recommendation algorithms, e.g., Collaborative Filtering, to recommend merchants to a target user. However due to the individual difference existing in user rating, rating values themselves do not give exact preferences of user. On the other hand, apart from user rating data, user reviews which convey accurate user preference information are inadequately considered in existing techniques. In this talk, we report our recent work addressing above two problems by 1) analyzing user reviews to discover user preferences in different aspects; and 2) leveraging the numeric order of ratings given by a user within a certain period to capture user real preference based on utility theory. We conduct experiments to evaluate the proposed approaches in terms of effectiveness, efficiency and cold-start using two real-world datasets. The experimental results show that our approaches outperform the state-of-the-art methods.

Attached file: