A Generative Adversarial Approach with Social Relationship for Recommender Systems

• DOI: 10.1145/3630138.3630424
• Publication Date: 2024-01-18T07:04:17Z
• Authors: Xiangxia Li; Kairui Chen; Yalan Zhou; Yanhong Chen
• Keywords: Mean reciprocal rank; Rank (graph theory); Learning to Rank; Generative adversarial network

With the explosive growth of data, personalized recommendation systems have become an essential component in digitally-driven society. However, one main challenges faced by traditional algorithms is difficulty obtaining accurate user preferences, especially for users with limited historical interaction data., which ultimately impacts performance these methods. To address this challenge, paper proposes a method, named STRGAN (Social trust relationships Generative Adversarial Network), it leverages advantages Networks (GANs) to tackle data sparsity problem, integrating ratings and social relationships. By incorporating both types information, aims improve accuracy quality recommendations provided users. Moreover, proposed model employs negative sampling techniques ensure that generated align real data. evaluate effectiveness STRGAN, extensive experiments were conducted on world dataset FilmTrust. The empirical results demonstrate outperforms other GAN-based models various evaluation metrics, such as precision, recall, normalized discounted cumulative gain (NDCG), mean reciprocal rank (MRR). offers robust efficient solution tasks. experiment support efficacy indicating its potential significantly real-world applications.