Construction of power marketing user knowledge graph based on $\text{BERT}+\text{BILSTM}+\text{CRF}$ model

In order to solve the fierce competition in the power market caused by the continuous improvement of the power consumption level of users, to reasonably deal with the contradiction between power supply and demand and to provide effective marketing strategies, the author proposes to construct the user knowledge graph based on the power marketing data. After preliminary data preprocessing, BERT (Pre-training of Deep Bidirectional Transformers for Language Understanding) model is selected for entity recognition. Bi-lstm (Bidirectional Long short-term Memory) model and CRF (Conditional Random Field) model are combined to extract the relationship, build triples and establish knowledge graph. The graph can clearly show the relationship between users and power load, address, price package, etc., which is conducive to relevant enterprises to put forward corresponding marketing or emergency measures, and provide theoretical support for optimization schemes such as “power demand side management” and “peak load shifting and valley filling” under the new situation.