Mobile Computing, Applications, and Services. 10th EAI International Conference, MobiCASE 2019, Hangzhou, China, June 14–15, 2019, Proceedings

Research Article

A Location and Intention Oriented Recommendation Method for Accuracy Enhancement over Big Data

Download
9 downloads
  • @INPROCEEDINGS{10.1007/978-3-030-28468-8_1,
        author={Wajid Rafique and Lianyong Qi and Zhili Zhou and Xuan Zhao and Wenda Tang and Wanchun Dou},
        title={A Location and Intention Oriented Recommendation Method for Accuracy Enhancement over Big Data},
        proceedings={Mobile Computing, Applications, and Services. 10th EAI International Conference, MobiCASE 2019, Hangzhou, China, June 14--15, 2019, Proceedings},
        proceedings_a={MOBICASE},
        year={2019},
        month={9},
        keywords={Intention-oriented recommendation Location-based clustering Spatial Performance improvement},
        doi={10.1007/978-3-030-28468-8_1}
    }
    
  • Wajid Rafique
    Lianyong Qi
    Zhili Zhou
    Xuan Zhao
    Wenda Tang
    Wanchun Dou
    Year: 2019
    A Location and Intention Oriented Recommendation Method for Accuracy Enhancement over Big Data
    MOBICASE
    Springer
    DOI: 10.1007/978-3-030-28468-8_1
Wajid Rafique,*, Lianyong Qi1, Zhili Zhou2, Xuan Zhao, Wenda Tang, Wanchun Dou,*
  • 1: Qufu Normal University
  • 2: Nanjing University of Information Science and Technology
*Contact email: rafiqwajid@smail.nju.edu.cn, douwc@nju.edu.cn

Abstract

Big data recommendation systems provide recommendations based on user history and optimize this process using feedback information. Recent developments in location-based social networks reveal that spatial properties of users greatly affect their opinion. Traditional location-aware recommendation systems do not consider user intentions to produce personalized recommendations. This paper proposes LIOR, a Location and Intention Oriented Recommendation method that uses spatial properties of users and their intentions to produce personalized recommendations. LIOR hierarchically employs user location and rating information to generate location-aware predictions, it then integrates user intentions to produce highly accurate recommendations. Extensive experimental evaluation performed on a real-world location-aware Movielens dataset demonstrates that LIOR provides exceptional performance on producing recommendations, it is highly scalable, and efficiently reduces the sparsity problem.