A Graph Neural Network Social Recommendation Algorithm Integrating the Multi-Head Attention Mechanism

Abstract

Collaborative filtering recommendation systems are facing the data sparsity problem associated with interaction data, and social recommendations introduce user social information to alleviate this problem. Existing social recommendation methods cannot express the user interaction interest and social influence deeply, which limits the recommendation performance of the system. To address this problem, in this paper we propose a graph neural network social recommendation algorithm integrating multi-head attention mechanism. First, based on the user-item interaction graph and social network graph, the graph neural network is used to learn the high-order relationship between users and items and deeply extract the latent features of users and items. In the process of learning user embedding vector representation based on the social network graph, the multi-head attention mechanism is introduced to increase the importance of friends with high influence. Then, we make rating predictions for the target users according to the learned user embedding vector representation and item embedding vector. The experimental results on the Epinions dataset show that the proposed method outperforms the existing methods in terms of both Recall and Normalized Discounted Cumulative Gain.

Social Network Community Detection to Deal with Gray-Sheep and Cold-Start Problems in Music Recommender Systems

“…We have pointed out in the previous section some of the classic ways of dealing with the problems that are the subject of this work and their main drawbacks. More recent works focus on incorporating social information to address a wide variety of issues that affect the quality of recommendations [25] or resort to newer techniques such as GNN-based methods [19,26], which are also not without weaknesses.…”

“…Recent research is moving toward the application of deep learning algorithms and Graph Neural Networks (GNN) [19], which provide good results when the volume of data is very large since they are endowed with a great ability to extract the complex relationships between users and items from large datasets. However, the target scenarios of this work are precisely those in which there is a scarcity of data, such as users with unusual tastes (gray sheep) as well as users with few interactions with the items (cold start).…”