Research On Tea Product Recommendation Based On Knowledge Graph Completion

China is a country with a long history of tea culture,and the sales of tea products are steadily increasing every year.In recent years,selling tea products on e-commerce platforms has gradually become a trend.The emergence of recommendation system helps users choose appropriate products according to their own preferences among many tea products.Through the processing of tea product related information,the recommendation system will recommend tea products that meet the needs of users to appropriate users,so as to improve the user’s consumption experience and increase the merchant’s sales revenue.Although the current tea product recommendation system has developed to some extent,there are still some problems,such as insufficient use of tea product information and poor recommendation performance.In this dissertation,the knowledge graph completion technology is applied to the research of tea product recommendation,and a tea product recommendation based on knowledge graph completion(TRKGC)is proposed.On the basis of constructing the knowledge graph of tea products,the model uses the emotional preference index extracted from user comments to improve the RSN knowledge graph technology;Using R-GCN to aggregate node neighborhood information,and by improving the jump mechanism of RSN,capture the information of nodes and edges on the graph path,so as to complete the knowledge graph of users and tea products.Finally,the feature similarity between the complement node and the original node is compared as the recommended evaluation index to recommend tea products.The main research work of this dissertation is as follows:(1)The knowledge graph of tea products is constructed and visualized.Firstly,the relevant tea product data of mainstream e-commerce platforms are obtained through web crawlers,and the user’s comments on tea products and other information are extracted;Then,word2 vec is used to obtain the comment feature vector and calculate the user’s emotional preference index to construct the knowledge atlas data set;On this basis,neo4 j is used to establish the graph database of tea product knowledge graph,and visual processing is carried out.(2)A tea product recommendation model based on knowledge graph completion technology(TRKGC)is proposed.Firstly,the R-GCN model is used to aggregate the neighborhood information of users and tea products to be recommended;Then,the emotional preference index extracted from user comments is used to enrich the edge information of relevant paths in the graph,so as to improve the RSN knowledge graph completion technology,and complete the missing triples through the RSN jump mechanism;Finally,the similarity between the complement node and the original node is compared as a recommendation index to realize the recommendation of tea products.(3)Evaluate the effectiveness of TRKGC model.The TRKGC model proposed in this dissertation is verified and evaluated by experiments from the aggregation range of neighborhood nodes and emotional preference index.The experimental results show that compared with other existing knowledge graph recommendation models,TRKGC has certain advantages in NGDC,MRR and other recommendation indexes.