Recommendation Model Based on Deep Separated Graph Neural Network

The recommendation system [1] can quickly find the information that users are interested in from the complex information and recommend it to the user, and it has been widely used in web-based services. However, the current mainstream recommendation models all have data sparseness [2] and imbalance problems. To address these problems, we propose a deep separation graph neural network recommendation model (DSGNR) based on the idea of graph information aggregation. In the experiment, we used the classic recommendation system data set to verify and analyze the proposed DSGNR model and compared it with the seven mainstream recommendation models. Experimental results show that the proposed algorithm is generally better than the current mainstream recommendation methods in terms of accuracy and efficiency.