电力事故案例的摘要提取方法研究  

Research on Abstracts Extraction Method of Power Accident Cases

作　　者：陈华千 CHEN Huaqian

机构地区：[1]广东电网东莞供电局,广东东莞523129

出　　处：《电力系统装备》2024年第11期165-168,共4页Electric Power System Equipment

摘　　要：随着电力系统的复杂性增加,电力事故案例的摘要提取对于事故分析和预防具有重要意义。文章从电力事故案例的结构、时间、空间及属性特征出发,采用文本挖掘技术提取关键词,利用自然语言处理技术进行实体和事件抽取,通过语义角色标注和信息融合构建事故描述,结合知识图谱和逻辑推理增强摘要的连贯性和完整性。最终通过模板和文本生成模型,自动化生成摘要,以期提高电力事故案例摘要的提取效率和质量,为电力安全管理提供科学依据。As the complexity of power system increases,abstract extraction of power accident cases is of great significance for accident analysis and prevention.In this paper,starting from the structural,temporal and spatial features of power accident cases,we adopt text mining technology to extract keywords,utilize natural language processing technology for entity and event extraction,and construct accident descriptions through semantic role labeling and information fusion.The coherence and completeness of the abstracts are enhanced by combining knowledge graph and logical reasoning,and finally the abstracts are automatically generated through templates and text generation models,with a view to improving the extraction efficiency and quality of the abstracts of electric power accident cases,and providing a scientific basis for electric power safety management.

关 键 词：电力事故 摘要提取 自然语言处理 知识图谱 事件抽取 

分 类 号：TM08[电气工程—电工理论与新技术]