Fast Recommender System Combining Global and Local Information : Construction of large-scale commodity information recommendation system

Large-scale recommender systems have important applications in online shopping and other fields.. In this paper, we propose a system that can balance the speed and accuracy of large-scale personalized recommender systems. To improve speed, we use GCN instead of the previous dense operation to obtain global information. At the same time, we use an attention mechanism so that various information can be fully integrated. The experimental results show that the method has obvious advantages in speed and accuracy, and the inference speed of only 109ms can achieve an accuracy of 83.1%.