Point-of-Interests Recommendation based on User-POI Coupling Relationships

• Published in: 2022 IEEE 8th International Conference on Cloud Computing and Intelligent Systems (CCIS) pp. 526 - 530
• Main Authors: Meng, Xiangfu, Zhang, Mingyang, Zhang, Xiaoyan, Fang, Jinfeng
• Format: Conference Proceeding
• Language: English
• Published: IEEE 26-11-2022
• Subjects: Cloud computing; Collaborative filtering; Computational modeling; Correlation; Couplings; Deep learning; Filtering; Neural network; Neural networks; POI recommendation
• DOI: 10.1109/CCIS57298.2022.10016353

This paper proposes a POI recommendation framework based on deep neural network. Firstly, we use k-means algorithm to cluster the POIs according to their geographical locations, so that POIs with high location proximity would be grouped into one category. Then, a convolutional neural network model is designed to study explicit correlations between users and POIs in aspects of respective features (such as correlations between users' age and POIs' location). Meanwhile, we use another neural network model to simulate matrix factorization method in machine learning. It can dig into implicit correlations between users and POIs according to users' ratings of POIs. Lastly, the explicit and implicit correlations between users and POIs are integrated to comprehensively represent user-POI coupling relationships. After, we put the user-POI coupling relationships into a fully connected network for recommendation. The model proposed in this paper has been evaluated on the Yelp dataset and the experimental results show that the model proposed in this paper achieves high recommendation accuracy.