Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Next and next new POI recommendation via latent behavior pattern inference
Li X., Han D., He J., Liao L., Wang M.  ACM Transactions on Information Systems 37 (4): 1-28, 2019. Type: Article
Date Reviewed: Jan 25 2022

The objective of a recommender system is to suggest relevant items or preferences to users. Similar to other recommendations, for example, for products, movies, or books, point-of-interest (POI) recommendations can help travelers and tourists find their next stop, like a restaurant or coffee shop, which can also benefit local businesses. Conventional next POI recommendation systems are based on users’ visiting history and check-ins, or use explicit data like ratings. Apparently, emerging location-based social networks (LBSNs) can be utilized to study user behaviors and patterns, which may increase the effectiveness of future recommendations.

Scattered historical record patterns are not sufficient to derive new results effectively. The analysis of certain data is often excluded when deriving conclusions. These kinds of challenges are addressed here by introducing two new areas of studies. The first area extends the existing global pattern distribution model (GPDM) [1] based on fixed pattern distribution. The paper proposes a new personalized pattern distribution model (PPDM) to discover unique pattern distributions for each user. This will help to deliver enhanced, personalized next POI recommendations.

The second area covers challenges related to next new POI recommendations. The paper presents novel next new POI research to deal with a real-world problem: new POI recommendations are not derived from the user’s visiting history. To evaluate the proposed models, GPDM and PPDM, various solutions are included for comparison (MF [2], PMF [3], FPMC-LR [4], PRME-G [5], GeoSoCa [6]).

A highlight of the study: the proposed models were extensively tested on three large-scale, commonly adopted real-world LBSN datasets, to show accurate recommendation regardless of time duration or distance. The evaluated effectiveness is clearly shown in the test results. The authors prove that the location category is an important factor for next or next new POI recommendations.

I would recommend this extensive study to data scientists who are involved in the analysis and design of recommendation systems.

Reviewer:  Brijendra Singh Review #: CR147403
1) He, J.; Li, X.; Liao, L.; Song, D.; Cheung, W. K. Inferring a personalized next point-of-interest recommendation model with latent behavior patterns. In Proc. of the 30th AAAI Conference on Artificial Intelligence ACM, 2016, 137–143.
2) Koren, Y.; Bell, R.; Volinsky, C. Matrix factorization techniques for recommender systems. Computer 42, (2009), 30–37.
3) Salakhutdinov, R.; Mnih, A. Probabilistic matrix factorization. In Proc. of the 20th International Conference on Neural Information Processing Systems ACM, 2007, 1257–1264.
4) Cheng, C.; Yang, H.; Lyu, M. R. ; King, I. Where you like to go next: successive point-of interest recommendation. In Proc. of the 23rd International Joint Conference on Artificial Intelligence AAAI Press, 2013, 2605–2611.
5) Feng, S. ; Li, X.; Zeng, Y.; Cong, G.; Meng Chee, Y.; Yuan, Q. Personalized ranking metric embedding for next new POI recommendation. In Proc. of the 24th International Conference on Artificial Intelligence AAAI Press, 2015, 2069–2075.
6) Zhang, J.-D.; Chow, C.-Y. GeoSoCa: exploiting geographical, social and categorical correlations for point-of-interest recommendations. In Proc. of the 38th International ACM SIGIR Conference on Research and Development in Information Retrieval ACM, 2015, 443–452.
Bookmark and Share
General (H.0 )
Data Mining (H.2.8 ... )
Information Systems Applications (H.4 )
Would you recommend this review?
Other reviews under "General": Date
Question answering in knowledge bases: a verification assisted model with iterative training
Zhang R., Wang Y., Mao Y., Huai J.  ACM Transactions on Information Systems 37(4): 1-26, 2019. Type: Article
May 10 2022
 Incorporating system-level objectives into recommender systems
Abdollahpouri H.  WWW 2019 (Companion Proceedings of the 2019 World Wide Web Conference, San Francisco, CA,  May 13-17, 2019) 2-6, 2019. Type: Proceedings
Jan 20 2022
Integrating runtime data with development data to monitor external quality: challenges from practice
Aghabayli A., Pfahl D., Martínez-Fernández S., Trendowicz A.  SQUADE 2019 (Proceedings of the 2nd ACM SIGSOFT International Workshop on Software Qualities and Their Dependencies, Tallinn, Estonia,  Aug 26, 2019) 20-26, 2019. Type: Proceedings
Jan 18 2022

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright © 2000-2022 ThinkLoud, Inc.
Terms of Use
| Privacy Policy