Application Recommendation based on Metagraphs: Combining Behavioral and Published Information

• DOI: 10.1109/compsac57700.2023.00039
• Publication Date: 2023-08-02T17:36:50Z
• Authors: Jinyi Wang; Tong Li; Hongyu Gao
• Keywords: Similarity (geometry); Conceptual model; Learning to Rank

Faced with so many mobile applications in the app store, users have difficulties finding their preferred applications. Existing studies do not comprehensively consider implicit feedback and thus combine behavioral information published together to make recommendations. This paper proposes a novel method recommend based on metagraph embedding using combination of information. Specifically, this constructed conceptual model combinations that could well portray Based model, six metagraphs are designed interpret multidimensional relationships between model. By random walking guided by each metagraph, series node sequences express neighborhood obtained. Finally, similarity apps is calculated embedded vector node, recommendations given user. real-world dataset, we evaluate performance our method. The experimental result shows outperforms existing models methods all metrics, which average F1-measure increases 19.21%, NDCG 4.99%.