A novel Recommendation Algorithm Based on Knowledge Graph Convolution Networks and LDA

Recommender systems exist with the aim of helping users discover personalized and useful content and products in a sea of information. Recommender systems have long faced challenges such as data sparsity and cold start. In this study, we propose a recommendation model based on Knowledge Graph Convolutional Network and LDA (KGCN-LDA), which first constructs a collaborative knowledge graph, and deep mines the auxiliary entity information in the collaborative knowledge graph using graph convolutional network(GCN). In the knowledge graph, GCN is used to deep mine the auxiliary entity information in the collaborative knowledge graph, while LDA mines the potential features in the semantic information, and finally, the user and entry embeddings obtained from the two parts are concatenated to generate the final recommendation results. In order to evaluate the effectiveness of the proposed model, we conducted experiments on a public dataset, and the results show that the proposed method significantly outperforms several existing methods in terms of precision, recall and F1.