Knowledge Graph-based User Portrait Construction for Electricity Enterprise Suppliers

Common user profile construction methods usually analyze user data from a single perspective to extract feature attributes and build user profiles. Since the correlation between different attribute features of users and the deviation of analysis perspectives are ignored, the constructed user portraits are low in granularity and poor in accuracy. To address the above problems, the study constructs the user portrait of electric power enterprise suppliers based on knowledge graph. The supplier knowledge graph is built by integrating and calculating the basic information data of suppliers. The basic information of power suppliers and the data in the supply process are used as the basic data sources, and after data processing, evaluation indexes are selected and the power suppliers are evaluated by using hierarchical analysis method. Combining the built knowledge map and evaluation results, a user profile labeling system is established to complete the construction of user portraits. The simulation test results show that the recall rate of the constructed user portrait can reach 98.87%, and the constructed user portrait has high granularity, which is better for practical application.