基于提示学习的工程施工经验智能引导及关键知识匹配方法

The intelligent guidance and key knowledge matching method for engineering construction experience based on prompt learning

摘要

工程施工是典型的知识密集型领域，经验知识是施工知识的重要来源，有效获取施工经验是企业的重要任务之一，然而此类经验知识具有碎片化、隐性化等特点，难以收集和利用。为此提出了施工经验智能引导及关键知识匹配方法。首先提出基于提示工程的文本分类技术用于施工经验知识内容定位，将专家经验文本划分成不同主题的文本块。在此基础上，通过对施工经验文本内容与施工规范条文特点的分析，提出利用正则表达式构建规则，利用规则匹配技术实现对专家经验知识要素的匹配分析，最终生成提示信息引导专家进一步思考和针对性回复。这种方法为工程施工经验的深度挖掘与利用提供了新的方法。

Abstract

Engineering construction is a typical knowledge-intensive field, where experiential knowledge is an important source of construction knowledge. Effectively acquiring construction experience is one of the key tasks for enterprises. However, such experiential knowledge is characterized by fragmentation and tacitness, making it difficult to collect and utilize. To address this, a method for intelligent guidance of construction experience and key knowledge matching is proposed. First, a prompt-based text classification technique is introduced to locate the content of construction experience knowledge by dividing expert experience texts into different thematic text blocks. Based on this, by analyzing the characteristics of construction experience texts and construction specification clauses, the study proposes a approach—regular expressions to construct rules. These rules are then utilized in matching analysis to align the elements of expert experiential knowledge, ultimately generating prompt information to guide experts toward further reflection and targeted responses. This method offers a new approach for the in-depth excavation and utilization of construction experience in engineering.

关键词

工程施工 / 提示学习 / 经验引导 / 文本分类 / 知识匹配

Key words

municipal engineering / prompt learning / experience guidance / text classification / knowledge matching

Author summay

杨华美(1986—), 女, 副教授, 博士, 主要从事水工混凝土性能研究。E-mail: yang_ [e-mail]

基金资助

中国长江电力股份有限公司科研项目资助(Z212302036)