吉林大学学报(信息科学版) ›› 2026, Vol. 44 ›› Issue (2): 467-474.

• • 上一篇    下一篇

基于主体语言模型的电力业务 AI 问答语义检索算法

林 巍, 李思韬, 林灶钢, 叶俊敏   

  1. 福建理工大学 计算机科学与数学学院, 福州 350118
  • 收稿日期:2023-11-24 出版日期:2026-04-14 发布日期:2026-04-15
  • 作者简介:林巍(1991— ),男,福州人,福建理工大学工程师, 主要从事自然语言处理、语音识别、信息安全等研究, ( Tel)86-17720757182(E-mail)wlin@fzu.edu.cn。
  • 基金资助:
    福建理工大学科研启动基金资助项目(GY-Z220209)

Semantic Retrieval Algorithm for AI Questions and Answers in Power Business Based on Subject Language Model

LIN Wei, LI Sitao, LIN Zaogang, YE Junmin   

  1. Department of Computer Science and Math, Fujian University of Technology, Fuzhou 350118, China
  • Received:2023-11-24 Online:2026-04-14 Published:2026-04-15

摘要:

由于电力业务涉及广泛而复杂的领域知识, 通常以非结构化形式存在, 包括电力系统运营、能源管理、电网规划等, 使电力业务智能化服务效率较低。为提高服务质量, 提出基于主体语言模型的电力业务 AI(Artificial Intelligence)问答语义检索算法。利用知识图谱建立电力业务知识库, 提供丰富的语义信息和知识存储; 采用基于词频-逆向文件频率(TF-IDF: Term Frequency-Inverse Document Frequency), 匹配电力业务 AI 问答语义, 在语义匹配的基础上建立主体语言模型(RWT-BERT: Retrieval-Augmented WaveNet Transformers), 利用该模型实现更加精确的电力业务 AI 问答的语义检索功能。 实验结果表明, 所提方法的检全率与检准率在 96% 以上, 平均倒数排名(MRR: Mean Reciprocal Rank)值最高达 94% , 具有较高的检索精度和效率。


关键词:

Abstract:

The power business involves a wide range of complex domain knowledge, usually existing in an unstructured form, including power system operation, energy management, power grid planning, etc. , which reduces the efficiency of intelligent services for power business. In order to improve service quality, a semantic retrieval algorithm for power business AI(Artificial Intelligence) Q&A based on the subject language model is proposed. A knowledge base for power business is established using a knowledge graph, providing rich semantic information and knowledge storage. Using TF-IDF(Term Frequency Inverse Document Frequency) to match the semantics of power business AI Q&A, a RWT BERT(Retrieval Augmented WaveNet Transformers) model is established based on semantic matching. This model is used to achieve more accurate semantic retrieval function for power business AI Q&A. The experimental results show that the proposed method has a recall and accuracy rate of over 96% , with an MRR(Mean Reciprocal Rank) value of up to 94% , indicating high retrieval accuracy and efficiency.

Key words:

中图分类号: 

  • TP311