Welcome to the InfoSci Platform
Reference Hub13
Fused Collaborative Filtering With User Preference, Geographical and Social Influence for Point of Interest Recommendation

Fused Collaborative Filtering With User Preference, Geographical and Social Influence for Point of Interest Recommendation

Jun Zeng, Feng Li, Xin He, Junhao Wen
Copyright: © 2019 |Volume: 16 |Issue: 4 |Pages: 13
ISSN: 1545-7362|EISSN: 1546-5004|EISBN13: 9781522564010|DOI: 10.4018/IJWSR.2019100103
Cite Article Cite Article

MLA

Zeng, Jun, et al. "Fused Collaborative Filtering With User Preference, Geographical and Social Influence for Point of Interest Recommendation." IJWSR vol.16, no.4 2019: pp.40-52. http://doi.org/10.4018/IJWSR.2019100103

APA

Zeng, J., Li, F., He, X., & Wen, J. (2019). Fused Collaborative Filtering With User Preference, Geographical and Social Influence for Point of Interest Recommendation. International Journal of Web Services Research (IJWSR), 16(4), 40-52. http://doi.org/10.4018/IJWSR.2019100103

Chicago

Zeng, Jun, et al. "Fused Collaborative Filtering With User Preference, Geographical and Social Influence for Point of Interest Recommendation," International Journal of Web Services Research (IJWSR) 16, no.4: 40-52. http://doi.org/10.4018/IJWSR.2019100103

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

Point of interest (POI) recommendation is a significant task in location-based social networks (LBSNs), e.g., Foursquare, Brightkite. It helps users explore the surroundings and help POI owners increase income. While several researches have been proposed for the recommendation services, it lacks integrated analysis on POI recommendation. In this article, the authors propose a unified recommendation framework, which fuses personalized user preference, geographical influence, and social reputation. The TF-IDF method is adopted to measure the interest level and contribution of locations when calculating the similarity between users. Geographical influence includes geographical distance and location popularity. The authors find friends in Brightkite share low common visited POIs. It means friends' interests may vary greatly. Instead of directly getting recommendations from so-called friends in LBSN, the users attain recommendation from others according to their reputation. Finally, experimental results on real-world dataset demonstrate that the proposed method performs much better than other recommendation methods.

Request Access

You do not own this content. Please login to recommend this title to your institution's librarian or purchase it from the IGI Global bookstore.