Distribution network fault equipment entity identification method based on BERT-BiLSTM-CRF model

The invention relates to a distribution network fault equipment entity identification method based on a BERT-BiLSTM-CRF model. A BERT-BiLSTM-CRF is used as a framework, a fault equipment entity recognition task is completed through multiple steps of original data processing, labeling mode design, model training and the like, a set of entity error correction and cleaning method based on the NLP technology is maintained at the same time, the labeling accuracy and recall rate are guaranteed, and the labeling quality is greatly improved. According to the method, a sequence labeled classic model BERT-BiLSTM-CRF is used, the semantic comprehension capability of a BERT large model is fully utilized, the large model is finely adjusted through a small amount of training data, and the entity recognition performance is greatly improved. According to the method, aiming at the problem of wrong marking in the entity marking task, according to the experience summarized in manual marking, the NLP technology is combined, so t