TY - GEN
T1 - Successive POI Recommendation with Category Transition and Temporal Influence
AU - Lin, I-Cheng
AU - Lu, Yi Shu
AU - Shih, Wen Yueh
AU - Huang, Jiun-Long
N1 - Publisher Copyright:
© 2018 IEEE.
PY - 2018/6/8
Y1 - 2018/6/8
N2 - With the popularization of mobile devices and wireless networks, people are able to share their experience on points of interest (POIs) in social networks through 'check-ins.' Therefore, the problem of successive POI recommendation has been proposed to recommend some POIs to users so that the users are likely to check in at these POIs in the near future. In this paper, we propose a two-phase method to solve the problem of successive POI recommendation. First, we utilize the Matrix Factorization technique to analyze the interaction of users and their sequential check-in behavior with time influence and POI categories, and select the candidate categories that the user will visit. Then, after removing those POIs not belonging to the candidate categories, we fuse user preferences, temporal influence and geographical influence together and finally recommend the POIs with high scores to users. The experimental results on a real check-in dataset show that our recommendation method is better than several state-of-the-art methods in terms of precision and recall.
AB - With the popularization of mobile devices and wireless networks, people are able to share their experience on points of interest (POIs) in social networks through 'check-ins.' Therefore, the problem of successive POI recommendation has been proposed to recommend some POIs to users so that the users are likely to check in at these POIs in the near future. In this paper, we propose a two-phase method to solve the problem of successive POI recommendation. First, we utilize the Matrix Factorization technique to analyze the interaction of users and their sequential check-in behavior with time influence and POI categories, and select the candidate categories that the user will visit. Then, after removing those POIs not belonging to the candidate categories, we fuse user preferences, temporal influence and geographical influence together and finally recommend the POIs with high scores to users. The experimental results on a real check-in dataset show that our recommendation method is better than several state-of-the-art methods in terms of precision and recall.
KW - Matrix factorization
KW - Recommendation
KW - Successive POI recommendation
UR - http://www.scopus.com/inward/record.url?scp=85055498376&partnerID=8YFLogxK
U2 - 10.1109/COMPSAC.2018.10203
DO - 10.1109/COMPSAC.2018.10203
M3 - Conference contribution
AN - SCOPUS:85055498376
T3 - Proceedings - International Computer Software and Applications Conference
SP - 57
EP - 62
BT - Proceedings - 2018 IEEE 42nd Annual Computer Software and Applications Conference, COMPSAC 2018
A2 - Demartini, Claudio
A2 - Reisman, Sorel
A2 - Liu, Ling
A2 - Tovar, Edmundo
A2 - Takakura, Hiroki
A2 - Yang, Ji-Jiang
A2 - Lung, Chung-Horng
A2 - Ahamed, Sheikh Iqbal
A2 - Hasan, Kamrul
A2 - Conte, Thomas
A2 - Nakamura, Motonori
A2 - Zhang, Zhiyong
A2 - Akiyama, Toyokazu
A2 - Claycomb, William
A2 - Cimato, Stelvio
PB - IEEE Computer Society
T2 - 42nd IEEE Computer Software and Applications Conference, COMPSAC 2018
Y2 - 23 July 2018 through 27 July 2018
ER -