基于多模型融合的电力运检命名实体识别  

Named Entity Recognition in Power Operation InspectionBased on Multi-model Fusion

作　　者：孙玉芹 肖静婷 王海超 SUN Yu-qin;XIAO Jing-ting;WANG Hai-chao(School of Mathematics and Physics,Shanghai University of Electric Power,Shanghai 201306,China)

机构地区：[1]上海电力大学数理学院,上海201306

出　　处：《科学技术与工程》2023年第36期15545-15552,共8页Science Technology and Engineering

基　　金：国家自然科学基金(11871377)。

摘　　要：为有效解决构建电力运检知识图谱的关键步骤之一的电力运检命名实体识别问题,通过构建一种基于Stacking多模型融合的隐马尔可夫-条件随机场-双向长短期记忆网络(hidden Markov-conditional random fields-bi-directional long short-term,HCB)模型方法研究了电力运检命名实体识别问题。HCB模型分为两层,第一层使用隐马尔可夫模型(hidden Markov model,HMM)、条件随机场(conditional random fields,CRF)和双向长短期记忆网络(bi-directional long short-term memory,Bi-LSTM)模型进行训练预测,再将预测结果输入第二层的CRF模型进行训练,经过双层模型训练预测得出最后的命名实体。结果表明:在电力运检命名实体识别问题上HCB模型的精确率、召回率及F1值等指标明显优于单模型以及其他的融合模型。可见HCB模型能有效解决电力运检命名实体识别问题。In order to effectively solve the power operation and inspection named entity identification problem,which is one of the key steps in building the knowledge graph of power operation and inspection,a hidden Markov-conditional random fields-bi-directional long short-term(HCB)model approach based on Stacking multi-model fusion was used to investigate the power operation and inspection named entity identification problem.HCB model was divided into two layers.The first layer used hidden Markov model(HMM),conditional random fields(CRF)and bi-directional long short-term memory(Bi-LSTM)model for training and prediction,and then input the prediction results into the second layer CRF model for training,and obtained the final named entity through the training of the two-layer model.The results show that the HCB model is significantly better than other models in terms of precision,recall rate and F 1 value on the identification of named entities for power operation and inspection.It is concluded that the HCB model can effectively solve the power operation and inspection named entity identification problem.

关 键 词：电力运检知识图谱 多模型融合 命名实体识别 隐马尔可夫-条件随机场-双向长短期记忆网络(HCB)模型 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]