图神经网络在软件安全需求分析中的应用探索.pdfVIP

2025年软件2025

第46卷第2期SOFTWAREVol.46,No.2

基金项目论文

图神经网络在软件安全需求分析中的应用探索

胡学锋

（晋中信息学院，山西晋中030800）

摘 要：为实现软件安全需求分析的自动化和精确化，本研究提出一种基于图神经网络的安全需求分析方法。通过构建

软件安全需求知识图谱，设计包含功能性安全需求、非功能性安全需求、安全威胁、安全资产和安全控制措施的节点体系，并

融合多头注意力机制设计图神经网络模型，实现需求依赖关系的自动识别。实验结果表明，该方法在NDCG@10、Kendalls

Tau等指标上均优于传统方法，专家一致性达到84.6%，为软件安全需求分析提供了有效的技术支持。

关键词：图神经网络；软件安全；需求分析

中图分类号：TP309;TP311.5 文献标识码：A DOI：10.3969/j.issn.1003-6970.2025.02.004

本文著录格式：胡学锋.图神经网络在软件安全需求分析中的应用探索[J].软件,2025,46(2):010-012

ExplorationoftheApplicationofGraphNeuralNetworksinSoftwareSecurity

RequirementsAnalysis

HUXuefeng

(JinzhongInformationCollege,JinzhongShanxi030800)

【Abstract】：In order to realize the automation and precision of software security requirement analysis,

the research proposes a security requirement analysis method based on graph neural network. By constructing

a software security requirements knowledge graph, designing a node system containing functional security

requirements, non-functional security requirements, security threats, security assets and security control measures,

and integrating the mechanism of multi-attention to design a graph neural network model to realize the automatic

identification of requirement dependencies. The experimental results show that the method outperforms the

traditional method in NDCG@10, Kendalls Tau and other indexes, and the expert consistency reaches 84.6%, which

provides effective technical support for software security requirement analysis.

【Keywords】：graph neural network;software security;requirement analysi