In order to solve the problem of sparse and unbalanced data in recommendation system, a separable graph neural recommendation model (SGNR) based on graph information aggregation was proposed. It models the group chara...
详细信息
In order to solve the problem of sparse and unbalanced data in recommendation system, a separable graph neural recommendation model (SGNR) based on graph information aggregation was proposed. It models the group characteristics of users and objects and extracts the collaborative characteristics of the group, which can effectively alleviate the problem of data sparsity and imbalance. In order to solve the problem that heterogeneous information network is difficult to construct, in the process of constructing graph neural network, the process of graph information collection is separated from the nonlinear transformation process of neural network, so that the modelling process is simpler. The experimental data is based on the data set of employee health information system, and the algorithm is verified and analysed by comparing various recommendation models. Experiments show that the SGNR method improves the efficiency and availability of the model, and the algorithm performance is greatly improved.
暂无评论