We found a match
Your institution may have rights to this item. Sign in to continue.
- Title
POI recommendation based on a multiple bipartite graph network model.
- Authors
Lang, Chen; Wang, Ze; He, Kaiming; Sun, Shimin
- Abstract
In location-based social network platforms, the point-of-interest(POI) recommendation is an essential function to serve users. The existing POI recommendation algorithms are rarely able to fully integrate various factors affecting the POI recommendation and cannot make dynamic recommendations to users over time. To address this issue, POI recommendation based on a multiple bipartite graph network model (MBR) is proposed. To reduce the overall complexity of the recommendation algorithm, we propose a clustering algorithm based on graph model, which determines the center of user clustering in the established user graph. An algorithm for finding the sparsest subgraph is built to cluster users who have social friendships or check in at similar POI, which significantly excludes more dissimilar users and makes the final recommendation more effective, as well as reducing the algorithm's complexity. To make more accurate recommendations to users, a large heterogeneous network of six weighted bipartite graphs is built based on the user clustering to describe the relations between users' social relationships, geographical locations of POI, and temporal information. The original Large-scale Information Network Embedding (LINE) model is too complex to be adopted for the learning of vertex embedding, thus it is optimized by negative sampling and Alias sampling methods and are fused with bipartite graphs, which accelerates the training speed. Finally, simulation experiments are conducted with the Gowalla dataset to verify the MBR algorithm, and the results show that the algorithm outperforms another three recommendation algorithms in terms of time awareness.
- Subjects
BIPARTITE graphs; GRAPH algorithms; SOCIAL networks; INFORMATION networks; SAMPLING methods
- Publication
Journal of Supercomputing, 2022, Vol 78, Issue 7, p9782
- ISSN
0920-8542
- Publication type
Article
- DOI
10.1007/s11227-021-04279-1