无线电工程2025，Vol.55Issue(5)：1105-1114,10.DOI:10.3969/j.issn.1003-3106.2025.05.023
基于BERT的混合电力信息报文自动提取技术
Hybrid Power Information Message Automatic Extraction Technology Based on BERT
贺云隆 1杨东华 1宋晓林 1张闯 2李佳燚 2刘鑫宇2
作者信息
- 1. 国网陕西省电力有限公司营销服务中心(计量中心),陕西西安 710100
- 2. 黑龙江省电工仪器仪表工程技术研究中心有限公司,黑龙江哈尔滨 150028
- 折叠
摘要
Abstract
In view of the fact that there is a lack of effective extraction technology for the large number of power information messages generated during the operation and maintenance of intelligent power systems,an hybrid power information message automatic extraction technology based on Bidirectional Encoder Representations from Transformers(BERT)is proposed,aiming at improving the analysis efficiency of unstructured fault text data in the power system.An automatic extraction model including BERT-SENet network for text classification and BERT-Bi-LSTM-CRF network for named entity recognition is studied and constructed.By integrating the tasks of text classification and named entity recognition,the model can effectively classify power fault texts automatically and extract key information.The experimental results show that the model is superior to the traditional methods in text classification and named entity recognition,which provides effective support for the informationization and intelligence of power system.Not only is the efficiency of power system fault handling improved,but also new ideas and methods for the automatic analysis of text data in similar fields are provided.关键词
智慧电网/电力故障文本/自然语言处理/文本分类/命名实体识别/双向编码器表示的Transformer模型Key words
smart grid/power fault text/natural language processing/text classification/named entity recognition/BERT分类
信息技术与安全科学引用本文复制引用
贺云隆,杨东华,宋晓林,张闯,李佳燚,刘鑫宇..基于BERT的混合电力信息报文自动提取技术[J].无线电工程,2025,55(5):1105-1114,10.
