基于大语言模型的函件文本分析方法

打开文本图片集

DOI:10.19981/j.CN23-1581/G3.2026.11.007

1 , 1 , 1 , 1 , 2

中图分类号:TP391.1

文献标志码:A

文章编号:2095-2945(2026)11-0029-07

Abstract: This paper proposes a Letter Text Analysis Method (LTAM) based on a large language model to address the issues of time-consuming and labor-intensive manual review of letter text information, as well as the risk of errors and omissions. This method first preprocesses the letter text, then converts it into text blocks through document splitting and stores them as vectors in the form of a local knowledge base. Secondly, in the face of problem query tasks, design prompt word templates, guide large language models to understand the text content of letters through prompt word methods and knowledge base construction methods, identify information that supports the integrity and effectiveness of files, and achieve the task of standardized and efficient management of letters in the early stage. The experiment shows that the letter text analysis method based on the big language model performs well in assisting reviewers in completing project reviews, and can quickly extract the summary of reply content and key information (opinions, dates, reply units, etc.).

Keywords: big language model; prompt word; text analysis; letter; knowledge base

在自然语言处理领域，人工智能技术取得显著进展，其中大语言模型得到广泛应用。（剩余14171字）