E-commerce Personalized Recommendation System Combined with Graph Convolutional Network and Big Data Optimization

An e-commerce personalized recommendation system is proposed, which combines Graph Convolution Network (GCN) and Large Data Optimization. The system constructs the graph structure of electronic commerce product and user, and makes use of graph convolution network to explore the potential relation of product and customer, and offer more precise personalized recommendations. Then, by using large data technique, the paper can clear and deal with the user's behavior data, reduce the disturbance of the noise data, and improve the data quality. To verify the validity of this approach, the paper simulates several e-commerce scenarios. The simulation results show that the recommendation system based on graph convolutional network has a significant improvement in the matching degree of user personalized needs, and the accuracy of system recommendations and user satisfaction are improved compared with traditional methods. The accuracy of system recommendations has increased by 15%, the user click rate has increased by 20%, and the response time of the system has been optimized, which can quickly provide personalized services in a real-time big data environment.