Product Recommendation Algorithm Based on Graph Neural Network

This paper proposes a product recommendation algorithm based on graph neural network to improve the accuracy and personalization of product recommendation. The main structure of the model is divided into three key steps: First, the feature of the user and the product is aggregated through the graph convolution operation, and the product information containing the user's preference is generated; Secondly, the attention mechanism is introduced to carry out weighted aggregation of commodity features under the commodity-user-commodity path, so as to capture the common commodity features that users are interested in. Finally, the fusion strategy is used to embed the above two features for effective integration to generate the final representation of the product, which is used to complete the downstream recommendation task. Experimental results show that the proposed method can significantly improve the performance of the recommendation system, especially in processing sparse interactive data and mining potential users' interests.